Figure 1 - uploaded by David Roy
Content may be subject to copyright.
Burn scar index change map values computed for a single hypothetical pixel over three consecutive ten day periods. A ® re occurs on day 25. The burn scar index change for the ® rst unburned period (days 11± 20) is a, for the second period (days 21± 30) is b, and for the third burned period (days 31± 40) is c. Note that the algorithm is applied to an additional unreported day of data at the end of each period.
Source publication
A multi-temporal burn scar detection algorithm designed for global application is described and demonstrated using 24-daily AVHRR images of an area of savanna burning near the Okavango Delta, Southern Africa. Thealgorithm is computationally simple, does not use fixed thresholds except to detect saturated AVHRR pixels, and incorporates a recent acti...
Context in source publication
Context 1
... burn scar index change map is computed as the maximum minus the minimum burn scar index value found for each co-located pixel in the diVerent orbits of data sensed over a given period. Figure 1 illustrates how a pixel which burns at some time during a ten day period will have a large burn scar index change value compared to the previous and subsequent periods when the pixel is unburned and burned respectively. ...
Similar publications
The archives of imagery and modeled data products derived from remote sensing programs with high temporal resolution provide powerful resources for characterizing inter- and intra-annual environmental dynamics. The impressive depth of available time-series from such missions (e.g., MODIS and AVHRR) affords new opportunities for improving data usabi...
While satellite-based monitoring of vegetation activity at the earth's surface is of vital importance for many eco-climatological applications, the degree of agreement among certain sensors and products providing estimates of the Normalized Difference Vegetation Index (NDVI) has been found to vary considerably. In order to assess the extent of such...
In this study, the emissions from forest fire in the Da Hinggan Mountains were estimated based on the TM, MODIS, and AVHRR satellite data, with spatial resolutions of 30 m, 500 m, and 1 km, respectively. In addition, the influences of different spatial resolutions on the emission estimation were quantitatively evaluated. Results showed that the dis...
Using logistic regression (LR) and artificial neural network (NN) algorithms, probabilities of presence/absence (p/a) of burned scars were calculated from post-fire Landsat 7 Enhanced Thematic Mapper plus (ETM+) images in mountainous areas of northern California. The discriminating power of six original TM bands (TM bands 1 through 5 and band 7) an...
At the present, the Cerrado (Brazilian savannas) is burned on about 20 to 30% of its area during the dry season mainly owing to anthropic causes. This work studied the susceptibility of the Cerrado vegetation to fire with respect to parameters linked to the anthropic activity. Location data for the fire spots obtained from the AVHRR/NOAA-12 channel...
Citations
... Widely used remote sensing indices for investigating vegetation, such as the Normalized Difference Vegetation Index (NDVI), have been tested for detecting burn severity. Roy et al. (1999) successfully applied NDVI to detect burn scars in a savanna area using a time series of remote sensing images. Further, Normalized Burn Ratio (NBR) is widely employed for assessing burn severity, with its performance being found to outperform single-band or NDVI approaches (Key et al., 2006). ...
Forest fires play a crucial role in resetting boreal ecosystems and steering ecological succession dynamics. However, the escalating impacts of global climate change are anticipated to increase the frequency, intensity, and size of wildfires, leading to significant economic, ecological, and social consequences. To effectively address fire risk and optimize post-fire management strategies, close monitoring, assessment, and understanding of the spatial heterogeneity of wildfires and their impacts are essential. Remote sensing, with its extensive historical records, provides a cost-effective means to examine wildfires. This study focuses on a significant wildfire event that occurred in May 2016 that made a substantial impact on Fort McMurray, Alberta, Canada. Using the Google Earth Engine (GEE) Platform, Landsat images time series covering pre- and post-fire (2015 to 2023), and land cover maps, we delineated the fire’s extent and conducted a comprehensive assessment of variations in burn severity and subsequent vegetation recovery. The Differenced Normalized Burn Ratio (dNBR) was calculated from Landsat images to measure burn extent, burn severity, and burn spatial variability. The Normalized Difference Vegetation Index (NDVI) was used for post-fire vegetation recovery analysis. Our findings reveal that 53.5% of the burn area experienced fire damage. Swamps and forests experienced the most intense burns (dNBR of 0.55 for swamps and 0.41 for forests) due to denser vegetation and biomass. Grasslands had moderate burn severity (dNBR of 0.281). In contrast, bogs, marshes, and fens showed lower dNBR values (0.15, 0.12, and −0.003), indicating low to no burns, likely due to their wetter conditions acting as natural firebreaks. NDVI changes indicate varying rates of vegetation recovery post-wildfire across different land cover types. In fen and marsh areas, NDVI was initially at 0.66 and 0.65 in 2015, dropped slightly in 2016, but rebounded by 2017, showing resilience. Swamps’ NDVI declined from 0.69 in 2015 to 0.46 in 2016, recovering to 0.72 by 2020. Grasslands’ NDVI dropped from 0.81 to 0.64 in 2016, recovering quickly to 0.80 by 2020. Forests’ NDVI decreased from 0.72 to 0.51 in 2016, with a gradual recovery to 0.67 by 2023, suggesting a slower recovery process. While NDVI values indicate a fast vegetation recovery for most land cover types, a deeper analysis suggests a transitional phase where past forests are now dominated by other vegetation types. The findings suggest that fire management strategies must integrate both immediate response and long-term recovery plans to ensure robust fire prevention and adequate rehabilitation resources for affected areas.
... As described in burned area products User Guide Version 3.0.1 of MODIS [45], the burned areas are characterized by deposits of charcoal and ash, removal of vegetation, and alteration of the vegetation structure [46]. To map the burned areas, the MODIS algorithm takes advantage of these spectral, temporal, and structural changes. ...
The estimation of the danger of fire is a way to quantify the potential or capacity of a fire to start, spread and cause damage. The relationships between the Canadian Fire Weather Index (FWI) System components and the monthly burned area as well as the number of active fire which has taken from Moderate Resolution Imaging Spectroradiometer (MODIS) Aqua/TERRA were investigated in 32 Guinean stations between 2003 and 2013. A statistical analysis based on a multi-linear regression model was used to estimate the skills of FWI components on the predictability of burned area and active fire. The goal of the detection algorithm is to identify "fire pixels" that contain one or more active burning fires at the time of the satellite overpass. This statistical analysis gave performances explaining between 16 to 79% of the variance for the burned areas and between 29 and 82% of the variance for the number of fires (P<0.0001) at lag 0. Respectively 16 to 79 % and 29 to 82 % of the variance of the burned areas and variance for the number of fires (P<0.0001) at lag0 can be explained based on the same statistical analysis. All the combinations used gave significant performances to predict the burned areas and active fire on the monthly timescale in all stations excepted Fria and Yomou where the predictability of the burned areas was not obvious. We obtained a significant correlation between the average over all of the stations of burned areas, active fires and FWI composites with percentage of variance between (75 to 84% and 29 to 77%) for active fires 35 and burned areas at lag0 respectively. While for burned area peak (January), the skill of the predictability remains significant only one month in advance, for the active fires, the model remains skilful 1 to 3 months in advance. The results show strong performances of the linear model in the sites with a homogeneous distribution of the vegetation cover like the case of the regions of Upper Guinea where one finds vast expanses of Savannah. Results also showed that active fires are more related to fire behavior indices while the burned areas are related to the fine fuel moisture codes. These outcomes have implications for seasonal forecasting of active fire events and burned areas based on FWI components, as significant predictability is found from 1 to 3 months and one month before respectively. These results also open up perspectives on the possibility of using forecasting models to project future and past events in order to better understand the long-term effects of bush fire on ecosystems, biodiversity and on the pollution of the environment.
... Yanmış alan verileri, bitki örtüsü olan bir yüzeyden kömür, kül ve çıplak toprak yüzeyine geçişi algılayarak yangından etkilenen alanları tanımlar (Roy, 1999). Yanmış alan haritalama esas olarak kalıcı olmayan bir arazi örtüsü değişikliğinin tespit edilmesi problemidir ve çeşitli algoritmalar uygulanarak konumsal/zamansal ölçekte farklı sonuçlar elde edilebilmektedir. ...
Bu çalışmada, Google Earth Engine (GEE) platformunda MODIS MCD64A1 ve European Space Agency (ESA) Fire_cci v. 5.1 (FireCCI51) küresel yanmış alan gözlem verileri kullanılarak Türkiye'de 2015-2020 yılları arasında meydana gelen yanmış alanların konumsal ve zamansal dağılımları belirlenmiştir. Elde edilen sonuçlar, Avrupa Orman Yangınları Bilgi Sistemi (EFFIS) verileriyle karşılaştırılarak yanmış alanların doğruluğu ve kapsamı değerlendirilmiştir. Çalışma kapsamında incelenen dönem boyunca, FireCCI51'in MCD64A1'e göre daha yüksek konumsal çözünürlüğe sahip olması ve küçük yanmış alanları daha hassas bir şekilde belirleme kapasitesine sahip olması daha fazla yanmış alan miktarları hesaplamasını sağlamıştır. Ayrıca, yanmış alanlardaki arazi örtüsü türlerinin belirlenmesinde MODIS MCD12Q1 arazi örtüsü sınıflandırma ürünü kullanılmıştır. 2015-2020 yılları arasında Türkiye'de meydana gelen yangınlar sonucunda yanmış bölgeler içerisinde, her iki veri setine göre de, arazi örtüsü türlerinde ekili alanlar en yüksek yanma oranına sahiptir. MODIS MCD64A1 verileri, ekili alanlardaki yanmış alan oranını % 88,93 ile % 91,80 arasında, FireCCI51 ise % 78,23 ile % 89,53 arasında belirlemiştir. Ülkemizde yanmış alanların konumsal dağılımı, özellikle Akdeniz, Ege ve Güneydoğu Anadolu bölgelerinde yoğunlaşmıştır.
... The produced information by (RS) is easily suitable for integrating in a Geographic Information System (GIS) that allows storing, processing, and displaying a lag amount of spatial information [12], as well as producing spatial analysis [13]. With respect to forest fires, (RS) provides information on environmental situation before, during, and after the fire [14]. At local to global scales, the satellite-based detection of fired areas has traditionally been carried out by the Advanced High-Resolution Radiometer (AVHRR) due to the high temporal resolution [15]. ...
A numerous wildfires have been recorded in the forests and grasslands of Kurdistan Region (KR) during recent years. Wildfires are a source of environmental, economic and social problems, human and population safety, and a real threat to human life at various scales and severities in different parts of the world. From an ecological point of view, fire is an important factor that plays a fundamental role in determining the diversity and dynamics of vegetation. Since (KR), located in north of Iraq, is almost the only area in the country where forests and vegetation seemed abundantly. The forests of (KR) play a vital and effective role in the vital, economic and touristic ecosystem of the region. Accordingly, it is very important to create and develop an accurate, rapid, and reliable spatial maps for understanding, interpreting and analyzing the causes and consequences of these fires. This study, which shows the possibility of fires in different areas of the study area, also aims to take precautionary measures in advance. Today, remote sensing data and technologies are among the most reliable mechanisms that provide spatial and temporal coverage of biomass burning, and determination of vegetation index (VI), without the complex, costly and stressful field procedures. Among these factors, the use of the normalized difference vegetation index (NDVI) was highly suggested as a useful tool for estimating the susceptibility of vegetation to fire. Accordingly, maps and data have been benefited from previous years, as this study attempted to prepare and develop a fire danger map by integrating satellite and field data for the Kurdistan Region. Times-series of MODerate resolution Imaging Spectroradiometer (MODIS) 250m, taken in 2010, were used for (NDVI) information layer. The developed map displayed a great consistency between the danger map, developed based on the 2010 image, and the site-recorded fire recorded from 2014 to 2015. The output map revealed that (RS/GIS) technology were of very high potential to be a valuable tool for managing, studying and controlling hazardous natural phenomena, such as wildfires, and reducing their risks and consequences. Distributed under the terms and conditions of the License 4.0 (CC BY-NC-ND 4.0) الملخص: ت ا حرائق تعتبر األخيرة. السنوات خالل كوردستان إقليم ومراعي غابات في الغابات حرائق من العديد تسجيل م ا مصدرً لغابات واالجتماعية واالقتصادية البيئية للمشاكل ، والسكان البشر وسالمة ، مس على البشرية للحياة ًا حقيقي ًا وتهديد في مختلفة وشدة تويات ا من مختلفة أجزاء بيئية نظر وجهة من لعالم. ، تنوع تحديد في ًا أساسي ا دورً يلعب ا ً مهم ً عامال الحريق يعتبر الغطاء وديناميكيات ألن ا نظرً النباتي. (KR) العراق شمال في الواقعة ، ال فيها بدت التي البالد في ًا تقريب الوحيدة المنطقة هي بكثرة. والنباتات غابات غابات تلعب (KR) ً حيوي ا دورً لذ ا ً وفق المنطقة. في والسياحي واالقتصادي الحيوي البيئي النظام في ً وفعاال ا لك ، إنشاء ًا جد المهم من الدراس هذه تهدف الحرائق. هذه ونتائج أسباب وتحليل وتفسير لفهم وموثوقة وسريعة دقيقة مكانية خرائط وتطوير ة ، تظهر التي من من مختلفة مناطق في حرائق نشوب احتمالية الدراسة طقة ، اليوم ا. ً مسبق االحترازية اإلجراءات اتخاذ إلى تعد وتقنيات بيانات الحيوية الكتلة لحرق وزمانية مكانية تغطية توفر التي موثوقية األكثر اآلليات بين من بعد عن االستشعار ، وتح الغطاء مؤشر ديد النباتي (VI) ، وال والمكلفة المعقدة الميدانية اإلجراءات دون العوامل هذه بين من مرهقة. ، استخدا اقتراح تم المعياري الفرق فهرس م النباتي للغطاء (NDVI) الخر من االستفادة تم فقد وعليه للحريق. النباتي الغطاء قابلية لتقدير مفيدة كأداة السنوات من والبيانات ائط السابقة ، الحريق مخاطر خريطة وتطوير إعداد الدراسة هذه حاولت حيث والميدان الفضائية البيانات دمج خالل من إلقليم ية المتوسطة الدقة ذو التصوير طيف لمقياس زمنية سلسلة استخدام تم كوردستان. MODIS)) 250 ا مترً ، التقاطها تم عام في 2010 ، معلومات لطبقة (NDVI). الخطر خريطة بين ا كبيرً ا ً اتساق المطورة الخريطة أظهرت تطويرها تم التي ص على ً بناء عام ورة 2010 من المسجل الموقع في المسجل والحريق 2014 إلى 2015 تقنية أن المخرجات خريطة كشفت. (RS / GIS) ذا كانت ت عليها والسيطرة ودراستها الخطرة الطبيعية الظواهر إلدارة قيمة أداة لتكون ًا جد عالية إمكانات ، ا حرائق مثل لغابات ، وتقليل وعواقب مخاطرها ها المفتاحية الكلمات : (GIS) ، بعد عن االستشعار (RS) ، النباتي الغطاء فهرس (VI) ، (NDVI) ، الغابات حرائق. پوخته : سروشتيه ئاگرکهوتنه له زۆر ژمارەيهکی ڕابردوودا سااڵنی ماوەی له له و دارستان كانی وه کوردستان ههرێمی ڕگاكانی تۆمارکراون و دانيشتووان و مرۆڤ سهالمهتی و کۆمهاڵيهتی و ئابووری و ژينگهيی کێشهی سهرچاوەی سروشتی ئاگری. ڕوا له جيهان. جياجياکانی ناوچه له جۆراوجۆر توندی و قهبارە به مرۆڤ ژيانی سهر بۆ ڕاستهقينهيه مهترسييهکی ژينگه نگهی يه وه دەگێڕێت بنهڕەتی ڕۆڵێکی که گرنگه فاکتهری ئاگر پێيهی بهو ڕووەک. دايناميکی و فرەچهشنی دياريکردنی له (KR) ک دەکهوێته ه به ڕووەک و دارستان که واڵتهکهدا له ناوچهيه تاکه نزيکهی عێراق، باکووری دارستانهکانی دەرکهوتووە. تێدا زۆری (KR) له دەگێڕن کاريگهر و گرنگ ڕۆڵێکی سيسته ژينگه می گهشتياری و ئابووری و ژيانی ی پ بهم ناوچهکهدا. نهخشهی گرنگه زۆر ێيه هۆکا شيکردنهوەی و لێکدانهوە و تێگهيشتن بۆ پێبدرێت پهرەی و بکرێت دروست متمانهپێکراو و خێرا و ورد شوێنی و ر ناوچهی جياوازەکانی ناوچه له ئاگرکهوتنهوە ئهگهری که توێژينهوەيه ئهم ئاگرکهوتنهوانه. ئهم دەرئهنجامهکانی لێکۆڵينهوەکه ههستک تهکنهلۆژياکانی و داتا ئهمڕۆ بگرێتهبهر. خۆپارێزی ڕێوشوێنی پێشوەخته ئهوەيه ئامانجيشی دەدات، نيشان دوورەوە له ردن ڕ پێوەرەکانی دياريکردنی و بايۆماس، سووتانی کاتی و فهزايی ڕووپۆشێکی که ميکانيزمهکانن متمانهترين جێی له ووەک (VI) مهي ڕێکارە بهبێ دەکهن، دابين بهکارهێن هۆکارانهدا، ئهم نێوان له سترێسييهکان. و تێچووناوی و ئاڵۆز دانييه پێوەرەکانی انی ئاسايی جياوازی ڕووەکی (NDVI) ئاگر بۆ ڕووەک ئامادەيی خهماڵندنی بۆ بهسوود ئامرازێکی وەک کرا پێشنيار زۆر پێيه بهم. پ بهو وەرگرتووە، پێشوو سااڵنی له سووديان داتاکان و نهخشه پهرەپ و ئامادەکردن ههوڵی توێژينهوەيه ئهم ێيهی نهخشهی ێدانی کات زنجيرە کوردستان. ههرێمی بۆ مهيدانی و دەستکرد مانگی داتاکانی يهکخستنی به دراوە ئاگرکهوتنهوە مهترسی ييهکانی MODerate resolution Imaging Spectroradiometer (MODIS) 250m ساڵی له که ، 2010 چي بۆ گيراوە، نه زانياری (NDVI) دا نيشان مهترسييهکان نهخشهی نێوان له گهورەی يهکدەنگييهکی پهرەپێدراو نهخشهی بهکارهێنرا. لهسهر که ، ساڵی وێنهی بنهمای 2010 ساڵی له که شوێنهکه تۆمارکراوی ئاگرکهوتنهوەی و پێدراوە، پهرەی 2014 تا 2015 تۆما رکراوە ل بهڕێوەبردن، بۆ بهنرخه ئامرازێکی ئاگرکهوت وەک مهترسيدارەکان، سروشتييه دياردە کۆنترۆڵکردنی و ێکۆڵينهوە نهوەی دەرئهنجامهکانيان و مهترسی کهمکردنهوەی و کێوييهکان،. كليله ووشه : GIS ، دوورەوە له ههستکردن (RS) ڕووەک پێوەرەکانی ، (VI) ، (NDVI) دارستانهکان، له ئاگرکهوتنهوە ، کێوی، ئاگری (MODIS).
... Data on the timing and extent of fires over 20 years were obtained from the MODIS burnt area product MCD64A collection 6 (Roy et al. 1999;Giglio et al. 2009Giglio et al. , 2016Giglio et al. , 2018. This monthly product identifies burn scars at a resolution of 500 by 500 m and provides an approximate (accurate to ±4 days) date of burning (Giglio et al. 2016). ...
Background
This paper presents an analysis of fire regimes in the poorly studied Angolan catchment of the Okavango Delta in Botswana. We used MODIS data to examine the frequency and seasonality of fires over 20 years (from 2000 to 2020) in three dominant vegetation types (miombo woodlands, open woodlands and grasslands, and short closed to open bushlands), and in areas where people were present, and where they were absent.
Results
The median fire return intervals for both open woodlands and grasslands and short bushlands were relatively short (1.9 and 2.2 years respectively). In miombo woodlands, fires were less frequent (median return periods of 4.5 years). Human population density had no discernible effect on the fire return intervals, but about 14% of the miombo woodlands experienced no fires over 20 years. Ongoing shifting cultivation within miombo woodlands has led to structural changes and the introduction of fire into this vegetation type where fires were rare or absent in the past. About 12% of the miombo did not burn during the period examined where people were present, whereas close to 20% of the sites remained unburnt where people were absent. This suggests that people did not change the fire return interval in any of the vegetation types studied, but that they altered the amount of the landscape that is flammable in miombo vegetation. Fires occurred between June and September, with a peak in the late dry season (August and September).
Conclusions
Historical research indicates that late dry-season fires are detrimental to miombo woodlands, and our analysis suggests that degradation in parts of the catchment has led to the introduction of fire to this previously fire-free and fire-sensitive vegetation type. Deforestation of miombo woodlands, and the consequent introduction of fire, is a cause for concern with respect to the ecological stability of the Okavango Delta. Managers should therefore aim to protect the remaining closed-canopy miombo stands from further clearing and to attempt to shift the timing of burns to the early dry season to reduce their intensity.
... The burn scars in the post-fire image are clearly noticeable. Burned areas are characterized by residues of charcoal and ash, removal of vegetation, and modification of the vegetation structure[83]. ...
Planted forests in South Africa have been affected by an increasing number of economically damaging fires over the past four decades. They constitute a major threat to the forestry industry and account for over 80% of the country’s commercial timber losses. Forest fires are more frequent and severe during the drier drought conditions that are typical in South Africa. For proper forest management, accurate detection and mapping of burned areas are required, yet the exercise is difficult to perform in the field because of time and expense. Now that ready-to-use satellite data are freely accessible in the cloud-based Google Earth Engine (GEE), in this study, we exploit the Sentinel-2-derived differenced normalized burned ratio (dNBR) to characterize burn severity areas, and also track carbon monoxide (CO) plumes using Sentinel-5 following a wildfire that broke over the southeastern coast of the Western Cape province in late October 2018. The results showed that 37.4% of the area was severely burned, and much of it occurred in forested land in the studied area. This was followed by 24.7% of the area that was burned at a moderate-high level. About 15.9% had moderate-low burned severity, whereas 21.9% was slightly burned. Random forests classifier was adopted to separate burned class from unburned and achieved an overall accuracy of over 97%. The most important variables in the classification included texture, NBR, and the NIR bands. The CO signal sharply increased during fire outbreaks and marked the intensity of black carbon over the affected area. Our study contributes to the understanding of forest fire in the dynamics over the Southern Cape forestry landscape. Furthermore, it also demonstrates the usefulness of Sentinel-5 for monitoring CO. Taken together, the Sentinel satellites and GEE offer an effective tool for mapping fires, even in data-poor countries.
... None of the ABI thermal wavelength bands were used because thermal wavelengths are better suited for the detection and characterization of actively burning fires as postfire emitted radiation varies rapidly in space and time (Justice et al., 2002;Wooster et al., 2003;Schmidt et al., 2012;Giglio et al., 2016). The middle-infrared (3-4 μm) includes reflective and emitted components and the reflective component has been shown to be useful for burned area mapping (Roy et al., 1999;Pereira 1999). However, derivation of the reflective component is complex as it requires middle-infrared emissivity retrieval which is difficult to undertake reliably on a systematic basis (Nerry et al., 1998;Tang and Li 2008) and so for this reason the 3.9 μm ABI was not used in this study. ...
The Advanced Baseline Imager (ABI) onboard the new generation of National Oceanic and Atmospheric Administration (NOAA) Geostationary Operational Environmental Satellite (GOES) series provides 10 min, multi-spectral, 500 m to 2 km observations, with significantly improved capabilities compared to the previous GOES sensors. For the first time GOES data are available with sufficient resolution to potentially enable burned area mapping using the reflective wavelength data. In this study GOES-16 ABI unburned and burned reflectance acquired over four extensive burns in the United States and South America were examined. Significant diurnal variations in the ABI reflectance due to surface reflectance anisotropy and changes in the position of the sun at each ABI observation time were observed, with unburned near-infrared (NIR) reflectance that varied by more than 50% over the day, and this variation was greater than the change in reflectance due to fire. The diurnal reflectance variation included locally increased backscatter reflectance acquired under near hot-spot conditions when the solar zenith was close to the view zenith. The spectral and diurnal temporal suitability of the ABI data for burned area mapping was assessed for the red, NIR and short wave infrared (SWIR) bands using the transformed divergence metric applied to the unburned and burned reflectance values for each ABI acquisition time. The separabilities varied diurnally but with no consistent view-solar geometry or time when the separabilities were consistently high among the four sites. The paper results demonstrate that use of ABI reflectance without consideration of, or correction, for surface anisotropy will be unreliable. The most suitable ABI bands to detect burned areas are the NIR, which provided high diurnal separabilities except in the early morning and late evening, and the normalized burn ratio (NBR), computed as the difference between the NIR and SWIR reflectance divided by their sum, that had more consistently high diurnal separability. Implications and recommendations for future ABI burned area mapping research are discussed.
... In order to obtain the area burned, the Moderate Resolution Imaging Spectroradiometer (MODIS) Burned Area product (MCD64A1, Collection 6), obtained from the University of Maryland's website, was used (Giglio et al. 2015). Burned area, which is characterized by deposits of charcoal/ash and changes in the structure of the vegetation, is mapped by a MODIS algorithm that takes advantage of the spectral, temporal, and structural changes in the land (Roy 1999;Roy et al. 2008Roy et al. , 2005Roy, Lewis, and Justice 2002). The approximate date of burning is detected at a resolution of 500 m by locating rapid changes in the daily reflectance time series data. ...
Emissions of ammonia (NH3), oxides of nitrogen (NOx; NO +NO2), and nitrous oxide (N2O) from biomass burning were quantified on a global scale for 2001 to 2015. On average biomass burning emissions at a global scale over the period were as follows: 4.53 ± 0.51 Tg NH3 year⁻¹, 14.65 ± 1.60 Tg NOx year⁻¹, and 0.97 ± 0.11 Tg N2O year⁻¹. Emissions were comparable to other emissions databases. Statistical regression models were developed to project NH3, NOx, and N2O emissions from biomass burning as a function of burn area. Two future climate scenarios (RCP 4.5 and RCP 8.5) were analyzed for 2050–2055 (“mid-century”) and 2090–2095 (“end of century”). Under the assumptions made in this study, the results indicate emissions of all species are projected to increase under both the RCP 4.5 and RCP 8.5 climate scenarios.
Implications: This manuscript quantifies emissions of NH3, NOx, and N2O on a global scale from biomass burning from 2001–2015 then creates regression models to predict emissions based on climate change. Because reactive nitrogen emissions have such an important role in the global nitrogen cycle, changes in these emissions could lead to a number of health and environmental impacts.
... Traditional BA approaches which combine surface reflectance imagery and AF data require the co-occurance of AF and the reflectance spectral signal to classify a pixel as a burned surface. In these methods, AF are used to derive statistics for burn classes or used as seed points in algotiyhms of regional growing (Libonati, et al. 2015;Roy 1999;Loboda, O'Neal, and Csiszar 2007). One of the major caveats of traditional classifiers is the underestimate of BA due to the accuracy of AFs products, usually caused by the satellite coverage (spatial/temporal), clouds, smoke, or sensor saturation (Holben 1986;Sousa, Pereira, and Silva 2003). ...
Adequate algorithms and spatial resolution remote sensing imagery still hamper a comprehensive representation of fire regime in regions characterized by small and sparse burnt scars, such as the southern Cerrado. The Visible Infrared Imaging Radiometer Suite (VIIRS) sensor launched in 2011 upgrades the spatial resolution (375 m) and gives continuity to the Earth long-term monitoring initiated by Advanced Very High-Resolution Radiometer (AVHRR) and Moderate Resolution Imaging Spectroradiometer (MODIS) sensors. Therefore, we developed a VIIRS 375 m burned area detection algorithm (VIIRS-SVM)
based on classification techniques specially adapted to regions with small and sparse scars aiming to improve the accuracy and reliability of burned area detection over the Brazilian Cerrado. For this purpose, (V, W) burnt index was adapted to VIIRS near-infrared and middle-infrared channels to enhance the burned area signal. Also, the One-Class Support Vector Machine classifier technique, a classification method based on machine learning, was used for burned area mapping. The proposed algorithm was applied over the Brazilian
Cerrado and evaluated against reference scars from 15 Landsat 8 OLI scenes, during the fire season of 2015, covering a large area with substantial burned area spatial variability. To analysis complement, we performed a comparison with the new released MCD64A1 collection-6 product. Results have demonstrated that operational automatic burned area mapping over the Cerrado is possible using VIIRS sensor capabilities. By taking advance of VIIRS 375 m coarse resolution, the proposed algorithm allows an enhancement of 25 % in discrimination of smaller and fragmented fire scars (50 ha and 250 ha), when compared to the MODIS-derived product.
... For example, algorithms using AVHRR data have been developed for Alaska, Canada, Siberia, Spain and North America (Kasischke and French, 1995;Fraser et al., 2000;Pu et al., 2007;Sukhinin et al., 2004;Fernández et al., 1997). Several algorithms have also been developed using AVHRR data for burned area estimates for Africa (Barbosa et al., 1999;Roy, 1999). Loboda et al. (2007) presented a regionally adaptive algorithm using MODIS data and applied it to North America and Siberia. ...
... Tansey et al. (2008) applied a temporal change index in which the daily re ectance was compared to a long-term running average composite. Roy (1999) also developed a multi-temporal algorithm based on the maximum change in a vegetation index between subsequent acquisitions. Often the length of the compositing window is allowed to vary as a function of the available observation sampling. ...
... Early examples of 'hybrid' algorithms were proposed by Roy (1999) and Fraser et al. (2000). Roy (1999) employed active re detections to re ne a statistical threshold on the di erence between two composites from AVHRR data. ...
Satellite-derived datasets recording the global extent of burned area (BA) have been generated in the past two decades. These products provide vital information to fire-related disciplines. The algorithms used to produce these products generally share a low level of methodological consistency, are highly empirical, and lack some of the sufficient features required to produce consistent and error characterised long-term data records (LTDRs). This thesis sets out to re-examine the quality of the algorithms used to produce these products with the aim of advancing the quality of future BA products. While BA products have been validated and inter-compared, little is known about the individual qualities and sensitivities of their respective algorithms. A novel sensitivity analysis of six global burned area algorithms provides information on the key sensitivities determining detection of burned area from the observations as well as highlighting the limitations of present algorithms. This analysis highlights the need for quantitative quality information within BA products through uncertainty characterisation. Best practice frameworks for uncertainty characterisation have been developed for many Essential Climate Variables (ECVs) but not yet burned area. The thesis then goes on to develop suitable methodologies for estimating uncertainties in present BA products and future products. Global uncertainties of 4--6\% are calculated for current BA products, and the necessary error propagation frameworks adapted to be amenable to binary ECV records such as burned area. The information gained from the sensitivity analysis and advancement of uncertainty characterisation provides a framework to prototype multi-sensor algorithms suitable for producing uncertainty characterised LTDR BA datasets. This involves the application of a spectrally invariant model of the burn signal suitable for mapping BA in a probabilistic manner from optical sensors to provide consistent estimates through the satellite record.
