Remote Sensing for Fire Management
Remote Sensing for Fire Management FOR 435
Popular in Course
Popular in Natural Resource Ecology And Mgmt
This 83 page Class Notes was uploaded by Ms. Alene Howell on Friday October 23, 2015. The Class Notes belongs to FOR 435 at University of Idaho taught by Staff in Fall. Since its upload, it has received 27 views. For similar materials see /class/227835/for-435-university-of-idaho in Natural Resource Ecology And Mgmt at University of Idaho.
Reviews for Remote Sensing for Fire Management
Report this Material
What is Karma?
Karma is the currency of StudySoup.
You can buy or earn more Karma at anytime and redeem it for class notes, study guides, flashcards, and more!
Date Created: 10/23/15
Remote Sensing of ActiveiFire and PostiFi re Effects Presentation 174 Energy and Energy Transfer Gaad Day quot1111 lemme Is gunned Energy and Energy Transfer 141 Understanding Energy in All its Forms What is Energy Physics De nition Energy is a quantity thalcan be converted into one form oranolher bulcannol be created ordeslroyed Types of Energy Kinetic Energy Energyassociated with an object in motion or lhe total workdcne to accelerate an object from rest to its total speed V2 mv Potential Energy A measure olthe potential for an object to interact Chemical Energy Energy derived from Chemical Interactions eg Fuel Let us rst consider what Energy actually is Energy is defined in physics as a quantity than be converted into one from or another but cannot be created or destroyed We are familiar with several different types of energy from Kinetic energy 7 associated with an object in motion and a function ofits mass and speed Chemical energy 7 associated with energy derived from chemical reactions like for example batteries or from gasoline Potential energy 7 associated with the potential of an object to interact with its surroundings For an example of potential energy consider when an object is dropped from a tall height In this case the total potential energy is calculated by the multiplication of objects mass by the height off the ground and the gravitational constant Ifthe height is tall enough and no energy is lost due to air resistance or other interactions this total potential energy would eventually be fully convertedinto kinetic energy 141 Understanding Energy in All its Forms Electromagnetic energy is generated by several mechanisms including changes in the energy levels of electrons decay of radioactive substances and the thermal motion of atoms and molecules Nuclear reactions within the sun produce a full spectrum of electromagnetic radiation Another from of energy is electromagnetic energy This type of energy is produced due to several processes such as changes in energy levels of electrons decay of radioactive substances and the thermal motion of atoms and molecules The fundamental source of all electromagnetic energy is the acceleration of electric charges soug m Sensor 1 Scmagsd amp rigwbed r Absorbed Atmospheric Limb Sounders Target mm mm Sun Atmospheric Occultation Sounder Let us now consider energy in terms of the remote sensing of the Earth Energy is emitted by the sun and travels through space until it interacts with the Earth s atmosphere There some of it is scattered back into space while some of it is absorbed or transmitted through the atmosphere About 43 of the incoming sun s energy makes it way through the atmosphere to reach the ground When the energy reaches the ground some of it is also absorbed and scattered and only about 7 of that energy will nd its way back through the atmosphere to be recorded by a satellite sensor When considering other forms of remote sensing other energy interactions are apparent For instance the two images on the right show examples of atmospheric remote sensing In the upper image the sensor evaluates how much solar radiation is scattered off the atmosphere to actually reach the sensor while the lower image relies on how much energy has been absorbed by the atmosphere Likewise in lidar remote sensing not all the laser pulses red at the ground make it back to the sensor Therefore an understanding of how energy interacts with its surroundings is important in the understanding of remotely sensed data 143 Energy Transfer Energy transfer describes the transit of energy due to a gradient Gradients could be due to various factors including temperature and pressure The general form ofthe energy transfer equation is Flux constant Final Gradient State 7 Initial Gradient State This equation is widely used in many aspects of science Energy transfer describes the tIansit of energy due to a gradient For instance in the case of heat energy transfer is the tIansit of energy due to a temperature difference auu i impi quot energy of all the particles within a substance So the faster the particles are moving the higher the mean kinetic energy andthe hotter an object will be The general form of the energy transfer equation is shown here It is simply a function of by how much the gradient has changed There are three main types of energy tIansfer mechanisms Namely Conduction Convection and Radiation 143 Energy Transfer Conduction An object transfers its kinetic energy ie Heat to another object byits molecules hitting the molecules making them move around of the coldero 39ec b t Hot Com Object Object 0 0 0 00 0 0 0 00 o o 000 0 0 o 0 o 000 00 0 0O 0 o0 00 0 00 0 Conductive Heat Flux kL Final Temperature 7 Initial Temperature Conduction refers to the transfer of heat across a medium To understand what happens in conduction let us consider a hot and a cold object that are initially set far apart quotquot quot 39 nhiem quot 39 hioh peed As such all the particles will have relatively high kinetic energies andtherefore the object will have a high temperature In contrast in the cool object the particles are moving very slowly andtherefore on average will have relatively low kinetic energies making the object have a low temperature Let us now place the hot and cold object beside each other Through chance encounters some of the fast particles from the surface of the hot object will collide with the slow moving particles on the surface of the cold object quoth 39 39 quot cause the lower 39 39 quot 39 39 faster and will also slightly slow down the faster particle Therefore this one collision will make the temperature of the hot object slightly drop while the temperature of the cold object will slightly rise This process is then repeated for all the other particles until the particles in both objects are traveling at approximately the same spee Energy Transfer Convection The kinetic energy of objects are moved from one location to another by physically moving the objects Heated Burned Surface I Temperature Gradient Convective Heat Flux h Final Temperature 7 Initial Temperature Hayman Five imenm R epun Convection refers to the transfer of energy by two mechanisms Namely diffusion as in conduction AND by the bulk motion of the object ie by physically moving all the randomly moving particles from one place to another To understand what happens in convection let us consider the example of a hot burned surface in a situation of no wind In this case as in the conduction slide the hot particles at the surface of the hot burned surface will through diffusion randomly transfer some of their kinetic energy to the molecules within the air that are close to the surface This will as before cause the air molecules to increase their kinetic energy and ultimately increase the air temperature Let us now add in some horizontal wind as shown in the figure on the left In addition to the random transfer of energy we now are effectively moving blocks of air from the left to the right The act of moving this air will transfer the faster moving air molecules slightly to the right of the interface area between the hot burned surface and the colder air As these moved air particles are hotter than the particles in new surrounding air they will continue through diffusion to pass on their energy to the slower moving colder air particles As in conduction this transfer will continue until a state of thermal equilibrium is reached The bulk motion of these particles are caused by two types of convection 143 Energy Transfer Radiation The transfer of energy via electromagnetic waves or photons is the only way the energy can be transferred within a vacuum ie in space between the sun and the Earth The StefanBoltzman Law E so a 567x1039E wattsmZIKA Radiation Heat Flux an Object T4 Background T4 Thermal radiation is a form of energy that is emitted by all matter above absolute zero The transfer of radiation occurs via electromagnetic waves and is importantly the only form of heat transfer that can occur in a vacuum A perfect emitter of radiation is called a blackbody The upper limit to how much energy an object can radiate at a given temperature is given by the StefanBoltzman law Where T is the absolute temperature of the objects as measuredin degrees Kelvin The term 8 e emissivity and denotes how efficient the object is at emitting radiation A value of 1 or 100 efficiency represents that the object is a blackbody The radiation heat flux equation is simply given as the difference in energy between that emitted by the object and that absorbed by its surroundings m FOR 434 Measuring Fire Effects Fire Effects on Soils Chad Hoffman and Zack Holden m2 My name is Zack Holden coinstructor for this course Many people associate severe re with catastrophic standdestroying forest fires But re can have profound effects on soils and the below ground processes that are critical for ecosystem structure and function In this lecture we will explore some of the major e ects that wildfires can have on soils Ecological Signi cance of Soils The soils we see today are the product of thousands of years of development In arid environments the Southwestern US Organic matter decompositionsoid buildup is very slow Soils mediate aboveground structurefunction 0 Structure OM Mineral structure Water balance porosity water retention Nutrient dynamics CN ratios 0 Microbial communities The soils that we see today are the product of thousands of years of organic matter decomposition and accumulation In many ecosystems particularly arid environments organic matter decomposition is very slow Soils mediate the aboveground characteristics that we see Soil structure is determined by organic matter content and mineral composition which in turn in uences it s water holding capacity These factors in uence nutrient dynamics and below ground microbial communities of the soil system Firg E ects Ovastory and Belowground Effects are important o Aboveground Firesdisturbance alter the biophysical characteristics of the ground surface increased solar insolation wind less canopy interception of precipitation Belowground Soil heating by fire will determine depth of burn amount of organic matter in soil consumed effects on soil microbes and Nutrient content of soil Disturbances like fire in uence both the aboveground and below ground characteristics of a site When we evaluate the effects of fire on soil characteristics both of these become important The most signi cant aboveground change that can occur postfire is the removal of overstory vegetation Think about what happens when we remove overstory trees More solar radiation reaches the ground and wind speed increases increasing soil temperatures and drying rates Without canopy interception of rainfall more rainfall reaches the ground which can increase soil erosion At the ground surface and below fiie eats soil directly which can volatilize nutrients destroy soil microbes and change the way water interacts with the soil surface Immdediate and shortatarm fire affects Fires potentially influence many of the biophysical characteristics of the ground surface Direct effects Indirect Effects Dead treesveg A radiation budget temp Soil heating A Water balance soil color lWind Exposed soil lOrganic matter It is useful to think of re effects on soils in terms of direct and indirect effects that together will determine the total effects of fire on soil characteristics Direct heating of the soil by fire kills plant roots and seeds exposes the mineral soil layer and volatilizes organic matter and nutrients in the soil These effects combined with the altered biophysical site characteristics that result from overstory vegetation removal determine the immediate and shortterm effects of fire on the belowground environment Fire Effects and postafiro environment 0 Combined with timing and intensity of precipitation events these effects will determine postfire erosion and longterm site vegetation response Vegetation Response Soil Erosion Response a 4 Together the direct effects and indirect effects of re combined with the frequency and intensity of rain events several years after the re will determine the soil erosion response of a site and hence it s longterm vegetation response Severity of Fire effects on soil depends on 0 Length of time fuel accumulates between fires and the amount of these accumulated fuels that are combusted during a fire Wells et al 1979 Properties of the fuels size flammability moisture content mineral content etc that are available for burning The effects of fuels on fire behavior during the ignition and combustion of these fuels Heat transfer in the soil during the combustion of aboveground fuels and surface organic layers Many factor can in uence the severity of re effects on soils These include the length of time between res and the resulting fuel accumulation The properties of fuels prior to combustion including size type and moisture content of fuels available for burning The effects of these fuels 0n resulting fire behavior in turn in uences how these fuels burn and hence how they heat soils Heat transfer in Soils Heat is transferred through soils by 4 mechanisms 0 Radiation o Convection Conduction molecular transfer 0 Vaporizationcondensation change in states that release energy as heat Heat is transferred into the soil via 4 mechanisms You should be familiar with basic mechanisms of heat transfer from introductory re behavior or re ecology courses Radiation and convection are the principle modes of heat transfer by which overstory vegetation is damaged during a wild re Active ames radiate heat that combined with hot air and gases can scorch or burn plant matter quickly The opposite is true of soil heating where the insulating effects of soil prevent rapid movement of heat downward into the soil It is estimated that only 10 to 15 percent of the heat generated by a wild re is transferred directly downward into the soil The key factor that determines how hot and how deeply that heat travels is duration or residence time Dry soil is a poor conductor but hot air and gases can move through dry soil via convection In moist soils the primary mechanism for heat transfer is conduction Think of a heavy metal pan placed on a hot stove Heat from the stove can t directly heat the cooking side of the pan It takes time for that heat to travel through the pan by conduction The process of heating is similar in soils where the heat from the re must travel layer by layer molecule by molecule deeper into the soil Another mechanism can also contribute to soil heating however Moist soil when heated releases heated gaseous water vapor that is capable of moving through small pores in the soil This super heated air can increase the rate of soil heating due to the high conductivity of water vapor Factors that Influence soil heat transfer 0 Soil moisture Soil physical properties particle size porosity Soil organic matter content Mineral soil is a very good insulator poor conductor Moisture content and organic matter content influence how well heat from fire is transferred deeper into soil Mineral soil is itself an excellent insulator and a poor conductor and the physical properties of the soil will strongly in uence the rate and depth of soil heating The size of soil particles the types of minerals in the soil and it s organic matter content in uence the structure and chemical composition of the soil which in turn determines how heat moves through the soil Water conducts heat more easily than mineral soil It is a much better conductor of heat than air Therefore given two areas of the same soil type one moist and one dry the moist soil will become heated Juulu quickly than the dry soil Temperature and Soil Effects The range of re effects on soil resources can be expected to vary directly with the depth of burn which is expected to vary as a function of the amount duff consumed and the degree of large woody fuel consumption As you can see from this graph soil temperatures quickly drop with increasing depth in the soil pro le However many of important ecological components in soil break down at fairly low temperatures Organic matter begins to break down at around 200 degrees centigrade and Nitrogen begins to volatilize at close to that emperature Seeds stored in the soil and plant roots are destroyed at even lowe temperatures FIRE SEVFRITY A i t o mama o a 16 Unbumed T a a s an a a g a H E a E 5 U Site Factors The severity of re effects on soils depends on duration of and depth of soil heating This gure illustrates the range of re severities for overstory and understory effects and the site level and environmental factors that in uence them At one extreme it is possible to have severe stand replacing crown re that leaves the understory and soil unburned At the other extreme in some ecosystems we see creeping surface and ground res that are extremely severe in terms of understory and soil effects but do little damage to the overstory vegetation These effects vary both in temperature and duration with crown res often burning rapidly and at extremely high temperatures and ground res burning at lower temperatures for hours or even days Fire Severity Past Fire Rccmc I 39nmn 11115 C Spn 39 t 1 s Runnng amp 1 Km lislk slamming 5pm 39nlmmcxl Canopy Height 8 Mass of Fine Fuel lfnbumud I In Depth 0139 Burn Site Factors Organic Soil Depth amp Course Wood The severity of above ground and below ground re effects should determine the recovery of vegetation post re Unburned areas can serve as refugia within burned areas Where re overstory re effects are severe plants must regenerate from seeds that survive in seedbanks or from seed sources that colonize off site Where re e ects on soils are severe for example where soil heating kills roots and eliminates seeds from seedbanks colonizing plants can come either from surviving overstory canopy plants the edges of less severely burned areas or from 01f site seed sources The combination of overstory and understory effects should determine the biophysical and chemical properties of the site and hence the longterm post re vegetation recovery 97 letter c 19650 c 1 Wnash GreenHve canopy Patchy burn a This photograph from the 2002 Hayman re in Colorado shows a low severity surface re Notice that much of the litter and ne fuels remain unburned and only some of the ground surface appears scorched 69 litter 20 soil 0 10 o ash 60 new litter Scorched needle cast In this moderately burned area more of the ground litter and fuels have been consumed Ash and exposed mineral soil are more abundant Notice that the brown needles most of which are scorched needles that have dropped from scorched trees cover much of the soil surface This is a key characteristic that separates moderate and severe postfire e ects Scorched needle cast can protect the soil surface from postfire rain splash and subsequent erosion 0 litter 69 soil 29 ash Nearly 100 tree mortality This site from the same re was very severely burned Almost all overstory vegetation was destroyed and on the ground litter fuel and duff was almost entirely consumed A lot of mineral soil is now exposed and black and white ash are much more abundant Assessing fire severity on soils estimate of energy radiated downward Integrated area under time temperature curve Appearance of the soil surface p Provides inference for undesired changes in site productivity sustainability biological diversity and watershed hydrologic response 0 Scale Matters wquot The severity of re effects on soils is determined primarily by the amount of heat energy that is radiated downward The only way to directly quantify this would be to calculate the res residence time and temperature at a particular location then integrate the area under this curve One of the only ways to to estimate this is by the appearance of the soil litter and duff We can draw inferences about changes in post re site characteristics and long term watershed responses based on what we observe in the eld It is important to keep in mind that scale is also important The size and location of severely burned patches will influence both the short and long term ecological responses to soil re Low depth of burn charred Litter charred or consumed Duff intact Woody debris is scorched to lightly Mineral soil unchanged Small twigs and some leaves remain on plants Ash color and Depth of Temp Temp Temp appearance burn Surface 12 cm 5 cm Black charred Light 100200 lt100 None litter A V Fire effects on soils can be estimated in the eld by evaluating depth of burn and can generally be related to surface temperatures and depth of soil heating Depth of burn is considered low when litter is charred or scorched but unconsumed and duff is intact Mineral soil will generally be left unchanged and temperatures below the soil surface will remain low Litter consumed Duff deeply charred or consumed Mineral soil visibly unchanged Ash light colored Twigs and branches mostly consumed Logs deeply charred Ash color and Depth of Temp Temp Temp appearance burquot Surface 12 cm 5 cm Bare soil Moderate 300 400 200300 4050 A K With moderate depth of burn small fuels are consumed and dulT is deeply charred or partially consumed More white ash is present on site indicating complete combustion of surface fuels and charring and comsumption of large diameter coarse woody debris is eVident Temperatures will be high in the upper layers of mineral soil and heat will begin to penetrate deeper into the soil High depth of burn Db cdnsur ned Surface mineral soil oxidized reddish or orange with visibly altered texture and structure Char layer dark color may extend several centimeters into mineral soil Logs deeply charred or consumed Almost no small diameter wood remains Kevin Ryan Fire Sciences Laboratory Missoula MT Ash color and Depth of Temp Temp Temp appearance bum Surface 12 cm 5 cm White ash red Severe 500750 350 450 gt100 soil at surface With high depth of burn nearly all surface fuels large and small will be consumed The mineral soil surface is exposed and shows signs of oxidation Organic matter beneath the soil surface will show signs of consumption and a black char layer is sometimes evident below the soil surface Temperatures at 5 centimeters and below can be quite high 1 1 Kevin Ryan Fire Sciences Laboratory Missoula MT Qquot The severity of re effects on soils can vary spatially with mosiac patterns evident at di erent scales In those photograph we see effects that vary from unburned areas with understory vegetation still present to complete litter and du consumption This variability can occur at a range of scales We can think of a variety of theoretical prefire characteristics that could in uence variability in re effects on soils Large coarse woody debris that burns in place for long periods of time often leaves patches of severely burned soil like the one in this picture Fuel bed characteristics and fuel moistures likely also determine how fuel burns and it s resulting effects on soil heating 20 Composite Burn Index ground estimates o Ocular postfire estimates of understory and ground effects Fine amp woody fuel and duff consumption change in soil color Estimated post fire erosion These estimates require no measurements and can be very subjective depending on your familiarity with the area and experience in making estimates E The eld estimates we ve discussed such as depth of burn and the appearance of postfire fuel and soil characteristics are incorporated into the Composite Burn Index that we discussed previously These include rating factors of the amount of ne and large woody fuel consumption and estimates of changes in postfire soil color and soil erosion There are currently no real quantifiable measures of the severity of postfire soil effects so keep in mind that such estimates can be quite subjective and will depend on familiarity with the area and prior experience in making these estimates Formation of Water Repellent Soils Volat39i39l ization of surface oFganic materialquot Litter and duff Temps 175 280 C Organic compounds cool then coat surface and subsurface soil particles Soils are left with reduced infiltration capacity and are more prone to erosion following rains Forest oor burned creating hydrophobic soil Courtesy Nazionai Park Service Another major effect that Wildfires can have on soils is the formation of hydrophobic or water repellent soils During the fire surface litter duff and overstory vegetation volatilize at temperatures above 175 degrees centigrade These compounds can coat soil particles in the upper layers of the soil These compounds can reduce the infiltration capacity of the soil reducing it s infiltration capacity This in turn can increase erosion rates and in some cases increases the likelihood of debris ows 22 i FirgEejts Testing for Water Repellency Water Drop Penetration Time WDPT test 0 Time to infiltration of a drop of water is measured Mini disk Infiltrometer 0 New instrument from Decagon Devices Inc Pullman WA Time to start of infiltration as well as the volume of water infiltrating in one minute is measured Several methods are now used to test for water repellency in soils An older method called the Water drop penetration test measures the amount of time need for a drop of water to penetrate and be absorbed by mineral soil A second method called the minidisc in ltrometer measures the vollume of water in ltrating the soil over a 1 minute period 23 quot F FirgE ects WDPT and Infiltrometer o WDPT and Infiltrometer tests are correlated and can be used independently or together to identify water repellent soils Relative infiltration rate provided by infiltrometer is more useful than time to infiltration of one water drop 0 Point measurements Both of these methods are point measurements and can be used together although they are both highly correlated with one another Developers of these techniques feel that the information provided by infiltrometer is more useful than the water drop penetration test 24 Water repellency by burn severity class Strengths Quantative measure W K momma Limitations 39 Can be highly V r r N g 7 variable i T Wain Difficult to scale up I to larger scales lw wva39e a a u kw These two methods are some of the only QUANTITATIVE measures available for detecting re effects in soils However as we ve seen there is a great deal of spatial variability in the severity of re effects on soils These measures can be used to detect water repellent soils However there is a great deal of variability in these data Measures can vary dramatically just a few feet from eachother In addition because these are point measures it is dif cult to scale these measures up to larger areas typical of most widlfires A F A FirgE ects 11019 Sensing 0 Extensive ground characterization of burn severity and soil water repellency Possibility of measuring soil and vegetation properties remotely o Predicting susceptibility of soil to post fire erosion amp weed detection 0 A large area can be covered quickly Relatively few ground samples are needed As we ve discussed in previous lectures there is need for remote sensing assessments of post re effects After large severe wild res it is important to identify areas with potential for erosion and debris ows Ongoing research is attempting to develop relationships between field measures and spectral measures from remote sensing platforms 26 Remote Sensing of ActivenFire and PostaFire Effects esentation 272 General Variations in Vegetation Spectra Good Day w g 221 Leaf Maturation In this lecture we will discuss how various natural processes alter the internal structure of leaves which in turn cause the reflectance properties of the leaves to change Before we begin let us again consider the crosssectional diagram of the typical leaf To recap In the visible part of the electromagnetic spectrum the pigments at the surface of the 221 Leaf Maturation Young Leaves Old Leaves As leaves become more mature ie age the structure of both the pigment and mesophyll layer changes causing noticeable changes in the reflectance of the vegetation The young leaves have compact mesophyll layers and are filled with small protoplasmic cells In contrast older leaves are filled with loosely packed large vacuolated cells in the mesophyll This structural difference makes older leaves more spongy and larger 221 Leaf Maturation MATURE LEAF mum LEAF REFLECTANCE PERCENT i 4 y so so LS 20 25 mm mm Source Colwell 1983 In this summary figure from Colwell 1983 we can see that the difference between a young and mature leaf for the same plant in very pronounced between 08 and 13 microns Essentially there is a difference of around 20 Such a difference if not accounted for could easily be mistaken for another process This example highlights the importance of accounting for phenology when doing remote sensing analysis Essentially if we acquire imagery at the same time each year then assuming there has be no extreme change in climate we can assume that on average the age of the plant leaves will be nearly the same As such if we notice a massive change of 20 between successive images then we can safely assume it is not due to leaf aging 222 Leaf Senescence Green Veg Senesced Vegetation Source Elvidge 1990 Further to aging as leaves start to senesce we actually see the structure of the leaf beginning to deteriorate As senescence proceeds the concentration of chlorophyll starch and protein decrease This results in a very notable shift in the red edge until it is entirely removed due to the loss of both the red and blue chlorophyll absorption features The subsequent yellowing and browning of leaves is due to the loss of the chlorophyll coupled with the emergence of carotene and 222 Leaf Senescence The senescence of Justicia brandegeana As the chlorophyll content decreases the pigment carotene which is yellow becomes dominant We see the visual effects of senescence clearly in these gures On the le we see the senescence resulting in the once green tree leaf becoming yellow as the dominance of the chlorophyll pigment is replaced by the carotene pigment On the right we see two successive Landsat images ofa rangeland The upper acquired in April and the lower in August Both are true color images and thus as our eyes would see the areas if we were ying overhead and looking down You can clearly see that the green re ectance has given way to yellowbrown colors 223 Vegetation Water Content Wm mm on pm am new A 3 8 Pam manac la Lulu usemmm w o 3 I 14 is m 20 WAVELENGTH IIIII rum Fig 3356 Relationship between leaf re ectance and water absorption in the 0 7 to 26 Lm wavelength region We will now consider the vegetation water content This feature in partly a function of age as younger leaves generally have lower water contents clue to the abundance of protoplasmic cells that do not exhibit high water storage capacities Water content can also be a function of season or due to drought Changes in the leaf water content also affects E 223 Vegetation Water Content REFLECTANCE PERCENT L i 50 I0 395 Source awell 198325 f 53 WAVELENGTH um In this summaIy figure from Colwell 1983 we can see the considerable differences in reflectance for a series of leaves from the same plant that are being progressively starved of water over the course of a week As water is being removed absorption due to water decreases causing a notable rise in reflectance at the 14 and 19 water absorption features Also notice that the reflectance between 08 and 13 microns also increases highlighting that although this area is mostly associated with the structure of the mesophyll layer the loss of water content also acts to increase the leaf reflectance in this part of the spectrum Fm Remote Sensing of ActiveiFire and PostiFire Effects Presentation 274 Spectral Indices Good Day m 241 Reflectance of GreenSenesced Vegetation and Soils quot7a REFLECTANCE VISIBLE NEAR 2 lNFRARED w lNIR 21 Let us again consider the spectral re ectance curve for green vegetation Again we can see from the lefthand image that the location of the 3rd and 4th bands of the Landsat satellite sensor are well placed to capture the properties of vegetation by highlighting both the re ectance in the red and nearinfrared spectral regions Therefore measurements in these bands can allow people to monitor processes such as water stress or senescence which as we can see in the image on the right causes the values in these bands to notable change E 241 Reflectance of GreenSenesced Vegetation and Soils Re ectance 0 030 050 070 090 110 13 150 170 190 210 230 25c wavelengthmm Let us now compare the spectral re ectance curves of green vegetation senesced vegetation and a typical black soil on the same gure where the location ofthe red and nearinfrared Landsat bands are highlighted by the red and gold columns respectively As you can see each of these three different surfaces has quite different red and nearinfrared re ectance values This clear distinction between these three common surface types in the red and nearinfrared bands has led to the development and application of simple models that take advantage of these marked differences in re ectance For example consider the ratio of the nearinfrared to the red re ectance which is called the Simple Ratio SR For green vegetation this would equal 455 9 For senesced vegetation we have 3525 14 And for this soil we would have 87 114 Therefore you can clearly see that the Simple Ratio can essentially be used to quickly and easily distinguish between these three different surface types This is one of the simplest examples of what we call a spectral index m 241 Re ectance of GreenSenesced Vegetation and Soils The Goals ofa Spectral Index Maximize the sensitivity of each surface of interest Normalize or reduce effects due to sun angle viewing angle the atmosphere topography instrument noise etc to allow consistent spatial and temporal comparisons Be linked to speci c and measurable surface processes Jensen 2000 Remote Sensing ofthe Environment Is Edition Prentice Hall Ideally a spectral index should only be used if it relates to a surface process via a physical process Spectral indices can take many forms from ratios like the simple ratio to linear or more complex band combinations For decades remote sensing scientists have used spectral indices to help them predict model or infer surface processes Although in this lecture we will only consider the use of spectral indices on unburned surfaces in the later lectures we will also consider how they can be used to map burned area and post fire effects Initially intrinsic indices were developed fmm cimnlo hand retina whinh hinhlinhtod tho amp 242 NDVI The Greenness Index The most Vllldely applied spectral Index I the Normalised Difference Vegetation Index NDVI NDVI pNIR T p1 pNIR TPR The most Widely applied spectral index is the Normalised Difference Vegetation Index or NDVI The NDVI index was first applied by Rouse et al 1974 and like the simple ratio makes use of the sharp contrast between the red and NIR reflectance values of green Vegetation The NDVI identifies the photosynthetic affinity or more simply the greenness of the Vegetation In cases of photo synthetically active Vegetation a low red reflectance is obsened along with very a 242 NDVI The Greenness Index Consider the following image of Moscow Mountain near Moscow Idaho The forested areas show clearly as dark green in this true color Landsat image Although we could use any standard image processing package to calculate the NDVI the following slide shows the necessary steps if you were using the ERDAS imagine software analysis package 242 NDVI The Greenness Index e m uassn er Mmslur ar omn AsE H lmag hummer lnpumle 39img Umpumle WEI Map me 5 an 25 nn 5 72375 nu Landsat w v I Shel2h m Untlgned a bu Select Funman lRR A suRn lRR 1 V2 index Funclmn band 6 hand a band Menu 3 Q am Am As you can see from this slide such packages normally have many standard spectral indices that can be easily selected and run in this way The calculation of the NDVI and other indices could equally easily be done in ARC or other GIS software packages 242 NDVI The Greenness Index 4am The output of the NDVI method is shown in this gure As you can see the heavily forested areas now appear bright white as they have values near 1 However to really see what is going on let us zoom in on the contents of the black box True Color Landsat Image B3 I32 Bl NIR Color Landsat Image B4 B3 B2 NDVI Image 1 Very Green The upper image is the true color composite Landsat image Ie what we would see if we were ying overhead and looking down on the scene In the middle gure the high nearinfrared values are now being highlighted in red high red values are being highlighted in green and high green values are being highlighted in blue This is an example of what we call a false color composite image This combination of bands is speci cally used to highlight green vegetation in red In this series of gures we can clearly see that the NDVI image has highlighted in bright values the forested areas The blackest area on this image is given by a lake near the center Other dark NDVI areas exist Speci cally we see areas that have some green re ectance Ieblue in the middle image but do not exhibit the high NIR re ectance This could correspond to meadows or clear cuts Remote Sensing of ActivenFire and PostnFire Effects Presentation 21 Spectral Properties of Unburned Vegetation Good Day 211 Reflectance of Green Vegetation Cell Let plglnlhlsi 51mm 3 WWWquot Emm rmm 1 ledmlliblmba an m PM WWW Watershsmpnnn mm W hsnrwnn mg i g 50 g 40 5 an n 2 14 15 19 2n 22 24 25 Wavelenglhlpm gtlt gt 311mm 1mm In order to use information of how fires affect the spectral properties of surfaces it is first necessary to understand the properties of surfaces that are not affected by fire Let us reconsider the spectral response curves of green vegetation as is shown in this gure As you may recall this diagram simply shows for each wavelength the amount of light that is re ected from a green tree leaf divided by the total amount of light that was incident on that leaf But questions you might ask are Do all green vegetation look the same And Why do green vegetation have curves that look like this Therefore In this lecture we will explain why the re ectance curves of green vegetation follow this general shape In later lectures we will then use this information to explain how these curves change to to fires and other natural processes 211 Reflectance of Green Vegetation Lets us consider the question of whether all green vegetation have the same general shape These plots show the spectral response functions of healthy green leaves for 9 different plant species This is from a comprehensive study of the spectral response curves of multiple green and senesced leaves as conducted by Elvidge 1990 As you can see although there are some difference the general shape of each healthy green leaf in essentially the same 211 Reflectance of Green Vegetation m cm PkaNrs 51mlme nmwm IALTOV WM Al coumnumo w REFLEEIANCE a mucrmce S mssamus e 7 04 06 M in 12 M La La 2u 22 24 wavumnm pl mums use mm IRAL gl la N R39 snow WAVE mrmw mum fo e Source 09120992 As all green healthy vegetation appears to have the same general spectral shape these re ectance properties must be due to general properties that are true for all leaves This diagram shows which of these general properties are dominant in different wavelength ranges Namely Leafpigments between 04 and 08 microns Cell structure between 08 andl3 microns Leaf Water Content 7 between 13 and 25 microns In the next few slides we will cover each of these properties in detail Notice that on this diagram the location of the Landsat satellite sensor bands have been highlighted As you can see this widely used sensor is able to capture information about each of these three dominant properties and thus is well suited to also monitor any changes that might occur Also note the two water absorption ands 7 these are wavelength ranges in which water in the air absorbs so much of the light that measurements in this areas are not use il in satellite remote sensing 211 Reflectance of Green Veetation 03 to 08 Microns 50 This region is governed by the absorption of the incoming EM radiation by pigments l Chlorophyll b Phycoerythrln Phymcyanm Chlorophyll a 39 l Refl ectance I Lquot ll Absorption is if 2 x 0 250 300 350 400 450 500 550 600 650 700 VloletBlue Green YellowRed 03 05 07 Wavelength nm Wavelength pm The reflectance of green tree leaves between 03 and 07 microns le the visible part of the electromagnetic spectrum is predominately governed by pigments in the leaves A common example of a pigment is chlorophyll which makes most vegetation appear green But you may ask how does it make the vegetation appear green To answer this consider the figure on the left Here you can see that the curves for chlorophyll a and b are shown in blue and green respectively Both these curves have 11 AA 211 Reflectance of Green Veetation 07 to 08 Microns The Red Edge 08 to 13 Microns The NIR Plateau 0 i i l 03 05 07 Wavelength pm The rededge is the sharp rise or step in green leaf reflectance seen between about 07 and 08 microns It occurs in part due to the near 95 absorption of chlorophyll in the red and the highly reflective nature of leaves between 08 and 12 microns The leaves are highly reflective in this range because of the internal structure of the vegetation Essentially in most vegetation whether healthy or senesced the spongy mesophyll layer scatters EM in the NIR This sharp contrast between the strong absorption 1 11 211 Reflectance of Green Vegetation 13 to 25 Microns 50 r This region is governed by the absorption of the incoming EM radiation by water and the thickness of the vegetation leaves Reflectance 1 9 21 23 2 Wavelength urn Between 13 and 25 microns the reflectance of green vegetation is mainly governed by the presence of water Most notably as shown in the right hand figure at 145 and 19 microns we can see the effects of the two dominant water absorption bands on the reflectance of the green leaves In leaves that are losing their water content due to water stress these dips in the green leaf reflectance become noticeably less pronounced In summary let us consider the following 211 Reflectance of Green Vegetation lransmrison specular reflection Uriuse ertchnn absummn cull39clc s 7 39 Pigments Palisade upper Eplrlcrmrs palisade parenchym04 spongy parcnchyma 4 mm crllular space T res39pirulian hula 2 sinmam In summary consider this crosssectional diagram of a leaf In the visible part of the electromagnetic spectrum the pigments at the surface greatly influence the perceived color of the vegetation The light with wavelengths above 08 microns transmits its way past these pigments and past the second layer of the leaf called the palisade but is finally reflected by the mesophyll layer As the first two layers don t absorb this light as was seen on its way in a considerable amount that is reflected by the mesophyll escapes the leaf As a result if vou viewed Remote Sensing of ActivenFire and PostaFire Effects Presentation 52 Mapping Vegetation Mortality and Recovery 11 Good Day 52 Spectral Meashres of Post Fire Effects Need immediate post fire measures that Relate to active re characteristics ie intensity amp Predict post fire effects ie severity When considering spectral measures of post re effects 7 clearly we could simply evaluate the changes in various spectral indices over time For instance ifwe followed the values of NDVI over time we could see how long it takes for a site to become as green or greener than the pre re condition This might give us information on the vegetation recovery 7 although it will remain dif cult to determine whether the green is due to saplings or grass le ideally we need the LIDAR height data to tell us what types ofvegetation it is from the height Therefore what information can we get about post re effects from spectral remote sensing To answer this lets us consider what we want remote sensing to tell us Since we ideally want information as soon as possible a er the re we want immediate post re measures that Relate to active fire characteristics ie intensity amp o Predict postfire effects ie severity 339 i i 4 Fire Intensity Fire Severity and Burn Severity Fr rn Jain Ti Piiiiod D Graham R 2004 Tongueued Wildfire 4 2226 After DeBano LP Neary DG Fioiiiott PF 1998 Fire s effects on ecosystems John Wiiey and Sons New York 333 pp The Fire Disturbance Continuum Tie Fire Di u mm Camillmm illusna39zs le our mpnnems includm in ailing re in Elwimlllnalll Elwimnmmml Emimmmi i39he hinlmical and 39 39 39 39 physicnl response m the minimum Seemum me mm m intuisily File diamum39su ns File menu um elleas hm combus lion prunes Firstalder re aliens Source of Confusion The Terms Fire Severity and Burn Severity are used inconsistently in the Remote Sensing literature Let us rst revisit the terms re intensity re severity and burn severity As we have discussed in this class these are confusing terms that exist on a temporal continuum VWtdkt2004Et39 The seVerltY concern an gaelrizg5eLeantile eta 2 Subjective amp Value Laden Term ANBR nonlinear asymptotic relationship with CBI that varies with sensor spatial resolution and environment El as a I unclinn 0 Hum Severity wu um um you 1200 V r Highlights need to evaluate alternative methods There are also numerous de nitions of severity and several negative connotations associated with the word as people typically associate the word severity with something bad Furthermore the current commonly used remote sensing severity index ithe Differenced Normalized Burn Ratio exhibits a nonlinear relationship with the commonly used eld measure of CBI that changes with the spatial resolution of the sensor used and the environment it is being applied in These unstable nonlinear relationships are a red ag in remote sensing as it means we can t easily scale our results from one analysis scale to the next In contrast a linear relationship is easy to scale This highlights the need to evaluate alternative methods either remotely or in the eld I Use Cover Fractions within a Pixel Direct Measure via Remote Sensing Comparable Measure via Field Methods Similar to traditional Green Brown Black Source Hudak Morgan Hardy et al One alternative approach is to measure the cover fractions within an area 7 whether it be quadrats on the ground or within a pixel These fractional cover measures are directly analogous to the traditional forest service measure of green brown and black which have been used for many years to assess post re effects Use Cover Fractions within a Pixel o Inherently Scalable 0 Use Existing 30m Immediate Postfire Landsat Data 0 Physically Related to Carbon and Water Processes Relations between percent cover measures ofbmn sevelity and carbon C and water 1120 cycles ET denotes evapotmnspiration Ecological Momes FireEiieors Reierehoes Linkages a c and H20 Cycles Tree sm vivalmonality Miller and W 2002 c accumulationET rates Litton et a1 2003 Tmmhore 2006 gare soil Gosfotli et a1 2005 Plant establishmentSoil respiration rates Reddehed soi1 Doeir and W 2005 In ltration Water r pe11ehcy and erosion Exposed litter Lewis et a1 2005 Plant establishmentWater i39epellency mm and Richardson 2000 Surface evaporation White ash Smith et al 2005b c volatilizatiouWater repelleucy Coarse woody debris Smith and Hudak 2005 c volatilizationerosion Importantly fractions are inherently scalable as we can report our areas in percentages and not values We can use existing post re Landsat data 7 which we have a lot of And several studies have shown that fractional cover measures are physically linked to several carbon and water processes Fr 5 A 1 Measuring Cover Fractions in a Pixel Standard and Simple Method Linear Spectral Unmixing Pixel Green Leaf Brown dry grass Black char We can break a pixel down into component fractions We make use of a simple remote sensing method that assumes that the re ectance of a given pixel is made up of a linear mixture given by a weighted proportion of the re ectance of green material a weighted proportion of the reflectance of brown material a weighted proportion of the re ectance of black material an error term This method is called linear spectral unmixing 7 as we are unmixing or finding the proportions of a linear mixture As we highlight here green could represent needles or tree leaves brown represents senesced grasses and black the char In this case study we are only considering the measure of the fractional char cover within a Landsat pixel as measured immediately postfire so within 16 days of the fire Note that in this method because of an error term it does remain possible to have fractions exceeding 100 but that the sum of all the fractions Le green brown and black within a pixel will sum to 100 39 A 7 Mapping Burned Areas Woodland Savannah Let us rst consider using this measure of the fractional char cover to map the area burned Consider this woodland savannah environment and the associated immediate post re Landsat image on the right 7 7 Setting a Char Fraction Threshold Accuracy 1 Kappa oxemu Amway u gt55 gt60 gt55 gt5u gt45 n r gt101 gt95 gtsu gt35 gtaa gt15 gt7 mm mmquot Ylwsslmlu 1 Landsat ETM 2 We rst need to de ne how much char needs to be present within a pixel for that pixel to be burnt As we can see from this graph if we assume that the pixel is burned if it contains a minimum of 50 char cover then our accuracy assessment in terms of both overall accuracy and kappa for this environment is near 100 Accumcy n KanDa a OvemllAcmmCy n gt101 gt95 ran gt35 win 15 7D gt55 gt50 gt55 gt5u gt45 Char mamquot Ylwaslmld gt50 Charcoal As you can see this map is fairly good 7 there are a few errors including several roads that have been highlighted r 33h Mapping Burned Area Comparison IKONOS 4m l Landsat 30m Let us now consider a neighboring region of both grasslands and woodlands In this comparison ofLandsat and IKONOS you can clearly make out the burned area Note the difference in spatial resolution between Landsat and IKONOS Also note that Landsat has 6 spectral bands while IKONOS only has 4 Mapping Burned Area Comparison i Landsat 30m Char Fraction gt50 v IKONOS 4m Supervised Classi cation If we run the char fraction method with the same 50 threshold on the Landsat image and compare it to a supervised classification of the IKONOS we can see that the char fraction appears to do a better job as it produced less coomission errors le the supervised classification appears to capture more noise probably due to roads or bare earth 12 39 A Mapping Burned Area Comparison y0984x Fo99 590155 r15 00 05 10 15 20 25 30 Lundmt ETM Am Buruml Ii 2 If we split each image into 16 regular squares and compare the area bumed we see that the char fraction method marginally underestimates the area burned in the IKONOS by about 16 which is a very low error Can We Use Char Fractions to Predict PostFire Effects Jasper Fire South Dakota started sperm 24 h Aug 2000 33800 ha Burned in 9 days Ponderosa Pine Forest Lentile et aI 20052006b 1Yr postfire Measures in 80 sites Landsat Image 14th Sept 2000 The question that now arises is can we use these measure of the char fraction to predict post re effects Let us not consider a pine savannah woodland in the Black Hills South Dakota This is the immediate post re image of the Jasper re that burned nearly 34000 ha in 9 days In this example we want to improve on that nonlinear relationship between dNBR and CBI Therefore rather than CBI which includes a large number of post re effects in its measure we will be looking at a few speci c post re effects that we would eXpect to be related to what the satellite sees T i 639 35 Also Measure Immediate PostFire ANBR Compare both ANBR and Char Fraction Cover to 1 yr Post Fire Field Canopy and Sub Canopy Measures Landsat 75 4 Char Fraction If we produce dNBR and the char fraction and place them side by side we can see that both capture the variation within the burned area The question is however how good is each approach at predicting canopy and subcanopy measures of the ecosystem condition one year post re Remember that dNBR gives us information about how band 4 and band 7 of the Landsat sensor changes due to the re This gives us information of the green vegetation from band 4 and the amount of soil or char on the ground from band 7 Whereas the char fraction tells us the actual quantity of char within a pixel Numerous agencies are currently applying dNBR and segmenting the images based on their values Le via thresholds into low moderate and high severity For instance in the middle gure these thresholds might cause all the red area to be classi ed as high and the orange as moderate The speci c thresholds oflow moderate and high severity are being evaluated in a range of different environments and are tweaked via the measurement of the composite burn index CB1 at each site After taking this class you should be able to ask critical questions on both the validity of this and similar methods and how this information is being used to evaluate post re vegetation and soil conditions L 4 1 Canopy Variables 1Yr Post Fire Immediate AN BR Live Tree 1000 230 200 we 310 am dNER MaraME Nam W Percentage Live Tree For this site let us now compare the dNBR as measured immediately postfire to the percentage live tree measured one year postfire As we can see 7 not the greatest of relationships This however does make sense 7 as dNBR measured immediately postfire tells us the immediate change in greenness from hand 4 between prefire and immediate postfire However several of these trees may have died within that subsequent year following the fire Also the immediate dNBR via band 7 gives us information on the soil and char cover which lyr postfire may be quite different due to regeneration as we see in the photograph l 1 4 E Canopy Variables 1Yr Post Fire Char Fraction Va Live Tree 060 0 7n u an we I on I 10 Char Ffacllol l Percentage Live Tree We can see that the char fraction measure does marginally better 7 probably because the amount of char on the ground might indirectly relate to the amount of trees that eventually will die 7 1 Sub Canopy Variables Immediate ANBR 40 Liner Organic Weighl grams dNBR NBRFRE NBRposr Litter Organic Weight gm2 l Let us now consider the subcanopy measure of the weight of organic litter as measured one year post re We can see a clear linear relationship between dNBR and this measure Therefore although signi cant scatter eXists this might be a Viable alternative to CB1 as it is a subcanopy measure that is telling us the quantity of post re fuels i A Sub Canopy Variables Char Fraction Liller Organic Weight grams N o O 60 070 080 090 1 00 110 Char Fraction Litter Organic Weight gm2 As you can see the char fraction measures again does slightly better A SubCanopy Variables Bole Scorch 10m 12m zoo 40B 600 sou dNBR NEEDEE r NBRDDQY Let us now consider the of the hole that is scorched which is a common measure included in the CBI As you can see there is no real relationship between dNBR and this measure This makes sense as dNBR is again giving us change in green and change in soil char and therefore there is no obvious physical reason why dNBR should relate to this measure A SubCanopy Variables Char Fraction Bole Scorch Bole Scorch surrogate of ame length 9 Intensity In contrast we see a strong relationship between the char fraction and this measure This potentially is because the char fraction is an indirect measure of the re intensity as a higher intensity re will combust more material and produce more c ar and where higher ame lengths are being translated into higher bole scorch values Therefore the char fraction is potentially an improved measure of the combined degree of litter consumption and subsequent litter recovery This is important as litter recovery has been shown to be particularly in uential in mitigating post re erosion Robichaud 2004 thus accurate prediction could assist managers with strategic treatment of severely burned areas Robichaud PR 2004 Post re Rehabilitation Treatments Are We Learning What Works Southwest Hydrology 5 2021 35 What s Potentially Happening with Bole Scorch Char Fraction 70 o P fraction surface and reports a value Within the pixel of70 35 7 L K What s Potentially Happening with Bole Scorch Char Fractlon 8500 As the re Intensity increases more orihe crowns WI combust This WI allow the satellite sensor to see more orihe charred surface and W111 cause the fraction to appear higher 35 What s Potentially Happening with Bole Scorch Char Fraction 10000 Win and report a value nearing 100
Are you sure you want to buy this material for
You're already Subscribed!
Looks like you've already subscribed to StudySoup, you won't need to purchase another subscription to get this material. To access this material simply click 'View Full Document'