Access the full text.
Sign up today, get DeepDyve free for 14 days.
Loading next page...
/lp/taylor-francis/a-google-earth-engine-approach-for-anthropogenic-forest-fire-SA0oB30al0
References
References for this paper are not available at this time. We will be adding them shortly, thank you for your patience.
- Publisher
- Taylor & Francis
- Copyright
- © 2023 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group on behalf of the International Water, Air & Soil Conservation Society(INWASCON).
- ISSN
- 2474-9508
- DOI
- 10.1080/24749508.2023.2165297
- Publisher site
- See Article on Publisher Site
Abstract
GEOLOGY, ECOLOGY, AND LANDSCAPES INWASCON https://doi.org/10.1080/24749508.2023.2165297 RESEARCH ARTICLE A google earth engine approach for anthropogenic forest fire assessment with remote sensing data in Rema-Kalenga wildlife sanctuary, Bangladesh Mohammed , Ariful Khan , Angana Kuri , Sohag Ahammed , Kazi Al Muqtadir Abir and Mohammed A.S. Arfin-Khan Department of Forestry and Environmental Science, Shahjalal University of Science and Technology, Sylhet, Bangladesh ABSTRACT ARTICLE HISTORY Received 8 August 2022 Due to global climate change, fire dynamics around the world are changing around the world Accepted 2 January 2023 affecting regions more severely that had negligible exposure to forest fire. In this study, we assessed the effect of anthropogenic forest fire using the Normalized Burn Ratio in Rema- KEYWORDS Kalenga Wildlife Sanctuary, Bangladesh while evaluating the relationship of this fire index with Forest fire; GEE; dNBR; different spectral indices, and Soil and Vegetation properties. A Correlation and a regression spectral indicies; climate analysis were used to examine the association of different variables with fire severity and One- change way analysis of variances (ANOVA) test was also performed to inspect the damage done by fire and the regrowth potential of the forest ecosystem. According to the result, the total burned area was not alarming. Also, the relationship of fire index with spectral indices, and soil and vegetation properties have been meaningful to understand the impact of the forest fire in Rema-Kalenga. Additionally, significant effects of the fire on the variation of the different properties were also observed. Even though the quantity of burned area in this forest was not significant, fire management in the forest conservation strategies in Bangladesh should be heeded and incorporated to avoid future disasters, keeping in mind the future climate change. Introduction 85% of forest fires were caused by humans (Mollicone et al., 2006). There are different factors that affect Forest fire has been a notorious disturbance phenom- anthropogenic fire understanding which is crucial for enon affecting different forest ecosystems throughout the prevention of fire. Among these, the presence of the world (Tyukavina et al., 2022). A fire incident can roads, human settlements and population density are be originated either due to anthropogenic or natural the noteworthy variables that have been studied to means. Due to global climate change, forest fire occur- better understand fire occurrence (CARDOSO et al., rence due to natural causes has soared over the past 2003; Korovin et al., 1996; Niklasson & Granström, years (Flannigan et al., 2006; Jolly et al., 2015; Nolan 2000; Wallenius et al., 2004). Unemployment, a socio- et al., 2021; Tyukavina et al., 2022) and it will like to economic indicator, has also been associated with fire get worse in the future as global climate condition will in different countries around the world (Catry et al., exacerbate favoring the global extreme fire weather 2009; Chuvieco et al., 2010; Sebastián-López et al., (Bowman et al., 2020). On the other hand, anthropo- 2008). Similarly, Arndt et al. (2013) also showed that genic climate change is considered the most talked human settlements and infrastructure developments about the human contribution to forest fire contribute notably to forest fire. (Flannigan et al., 2013). Human involvement with Burn severity is a commonly used term that fire occurrence is more pronounced in terms of direct describes the extent of damage caused by fire. Forest involvement. According to Ganteaume et al. (2013), in fire burn severity can be defined as the extent of loss of the Mediterranean countries, for a large number of fire vegetation and soil properties resulting due to the occurrences, human activity was responsible from the occurrence of fire (Taylor et al., 2014). To better period 2006–2010. Another study (Reineking et al., understand the effect of fire on the overall forest 2010), spanning over a 37-year period, on the fire- ecosystem, the relationship of fire severity with the prone areas of Switzerland indicated that anthropo- biophysical properties of the forest should be evalu- genic activity was the main reason for fire ignition ated. Fire severity is largely governed by fuel presence, even when other fires igniting conditions were not terrain characteristics, and weather which help to present on the study area. Forest fires in Canada determine the level of severity the area is likely to (from 1977 to 1999) were mostly people-induced as suffer (Parks et al., 2018; Roberts et al., 2003). presented by (Wotton et al., 2003) and another study Different studies investigated the correlation between on the Siberian boreal forest shows that more than CONTACT Mohammed A.S. Arfin-Khan mohammed.sust.07@gmail.com Department of Forestry and Environmental Science, Shahjalal University of Science and Technology, Sylhet, Bangladesh © 2023 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group on behalf of the International Water, Air & Soil Conservation Society(INWASCON). This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. 2 MOHAMMED ET AL. field-level biophysical data and different fire severity Table 1. Overview of previous research on forest degradation indices (Dindaroglu et al., 2021; Moody et al., 2016; caused by anthropogenic disturbances in different forests of Bangladesh. Rozario et al., 2018; Smith et al., 2021) and the results Studied Parameters of of these studies helped produce models that predict Anthropogenic the fire behavior in a forest ecosystem (De Santis & disturbances Study Area References Chuvieco, 2007; Dindaroglu et al., 2021; Smith et al., Deforestation or Cox’s Bazar – Teknaf (Hasan et al., 2021), illegal tree cutting Peninsula (Sobuj & 2021; Yin et al., 2020). Besides, apart from Normalized Rahman, 2011) Burn Ratio indices – Normalized Burn Ratio (NBR) Raghunandan Hills (Redowan et al., Reserve forest 2020), (Redowan and differenced Normalized Burn Ratio (dNBR) - et al., 2019) there are different spectral indices that showed suffi- Rema-Kalenga Wildlife (Islam et al., 2017) Sanctuary cient accuracy in explaining and mapping burned area Sal Forests (Rahman et al., (Dindaroglu et al., 2021; Hao et al., 2022; Mallinis 2010) et al., 2018; Sobrino et al., 2019; Veraverbeke et al., Agriculture Fasiakhali Wildlife (Rana & Hafizur establishment Sanctuary Rahman, 2021) 2011). Therefore, assessments of the burned area using Sal Forests (Rahman et al., these spectral indices have provided satisfactory solu- 2010) Satchari National Park (Mukul et al., 2017) tions for finding an alternative to field investigation of Grazing Ratargul (Humayun-Bin- burned area mapping of different temporal and spatial Akram & Masum, 2020), scales. (Sobuj & Forest fire is an issue faced by most of the vegetative Rahman, 2011), (Islam et al., areas around the globe except for regions near the 2017) poles (Mouillot & Field, 2005). In most cases, these Tourism Ratargul (Humayun-Bin- fire occurrences originated from either anthropogenic Akram & Masum, 2020) and natural sources. However, in most South Asian Fire Bangladesh Sundarbans (Hasnat et al., countries, the origin of these fire events is mostly Mangrove Forest, 2018) Chittgong Hill Tracts anthropogenic (Reddy et al., 2019). Tropical forest South Asian countries (Reddy et al., 2019) regions in South Asian countries are among the most Bangladesh (Ommi et al., 2017) Overexploitation Bhanugach Reserved (Halim et al., 2008) diverse ecosystems in the world, but there are a limited Forest amount of research done on the issues of forest fire in Ratargul (Humayun-Bin- Akram & Masum, this area (Reddy et al., 2019) which makes it apparent 2020) that a significant research gap exists on the under- Sal Forests (Rahman et al., standing of resilience capacity of these ecosystems 2010) Rema-Kalenga Wildlife (Islam et al., 2017) under global climate change scenarios. Sanctuary Due to the limited bodies of research, regarding forest fire, South Asian countries like Bangladesh have poor fire management facilities in their forest For conducting reforestation or afforestation activ- sectors (Hasnat et al., 2018; Ommi et al., 2017; Reddy ities after any kind of forest fire incident, especially in et al., 2019). Other than fire, there were some research a country like Bangladesh where forest fire is not heeded works done on other forms of human disturbances with such concern, a study showing spatial distribution (Deforestation, Grazing, Overexploitation, and magnitude of fire-affected areas and quantification Agriculture expansion, Tourism) in different forests of damage level is imperative. The main objective of this of Bangladesh which are pointed out in Table 1. In the study is to identify the distribution and measurement of forests of Bangladesh, naturally occurring fire is different burn severity classes in the Rema-Kalenga almost absent while man-made fire occurs regularly Wildlife Sanctuary (RKWS). In addition, another objec- before the monsoon season for the purpose of clearing tive is to assess the relationship of different spectral land for shifting cultivation, deforestation, firewood indices, vegetational and soil properties with burn collection, etc. (Hasnat et al., 2018; Mohammed indices, as well as to determine whether spectral indices, Bhuiyan et al., 2022). Under the future climate change vegetational and soil properties differed significantly scenarios, disaster anomalies that were once rare in among the burn severity classes. To achieve these objec- particular countries might start to occur in those tives, we derived all of the spectral indices from the regions frequently. For instance, global fire dynamics different satellite datasets (Sentinel and Landsat). We will change under future climate change projections also conducted a field-level survey to collect vegetation (Bowman et al., 2020; Flannigan et al., 2013) and and soil data along with reference data for burn indices. Bangladesh is more likely to be subjected to forest According to our knowledge, this study is the first to fires on a regular basis as natural and anthropogenic assess the burn severity of forest fires in Bangladesh by causes of fire will become more abundant combining remote sensing and field data. By conduct- (Chowdhury & Ndiaye, 2017). ing this survey, we hope to derive spatial and GEOLOGY, ECOLOGY, AND LANDSCAPES 3 quantitative depiction of the burned area along with This sanctuary’s terrain consists of undulating hills understanding the relationship between burn indices of varying heights and tiny, low-lying valleys. The with other spectral indices and different biophysical highest hilltop is approximately 67 meters above sea data, and the difference of these spectral and biophysical level (Sobuj & Rahman, 2011). Several hill ridges are datasets in terms of burn classes. extending in various directions. This region’s soil range from clay to sandy loam and have a low pH. In certain instances, the soil texture consists of a mixture of yellowish red sandy clay and grains of magnified Materials and methods iron ore. The average temperature in the region ranges Study area from a low of 27°C in February to a high of 37°C in June, with the average monthly humidity ranging Rema-Kalenga Wildlife Sanctuary (RKWS) is a tropical from 74% in March to 89% in July and average annual evergreen and semi-evergreen forest (Khan et al., 2002) precipitation of around 4,000 millimeters (Sharma, in the north-eastern area of Bangladesh (Figure 1a-c) in 2006). The texture is composed of a mixture of yellow- Chunarughat Upazila of Habiganj district in Sylhet ish-red sandy clay and granules of magnificent iron division, between 24°06“and 24°14” north latitude and ore. There are Bengali and Tripura ethnic minority 91°34“and 91°41” east longitude (Chowdhury & Koike, settlements (about 36 villages) both inside and outside 2010). RKWS was proclaimed formally for the first time of RKWS. Within the sanctuary, registered forest vil- in 1981, with an area of 1,095 hectares, then increased in lagers are granted certain privileges, including author- 1996 to encompass an area of 1,780 hectares ization to reside on forest department’s property and (Chowdhury & Koike, 2010). It is a portion of the cultivate low-lying land. In exchange, they contribute Tarap Hill Reserve, the biggest highland reserve in to crop management and forest conservation (Bhuiyan Bangladesh, and is situated along the Indo-Burma et al., 2021). Biodiversity Hotspot in the southeast of the nation (Sarker et al., 2013). It is part of the 9b-Sylhet Hills bio-ecological zone. It is bordered by reserve forests Satellite datasets on the majority of its northern and western boundaries, the Rema Tea Estate on its southwestern boundary, For the analyses of burned area, surface reflectance India (Tripura State) on its southern and eastern datasets of both, Sentinel-2A and Landsat 8, were used boundaries, and Khas land (owned by the national which were accessed through the Google Earth Engine government and managed by the district land adminis- (GEE) cloud computing platform. Both of these ima- tration) on a small portion of its northern border geries are already atmospherically and geometrically (Sharma, 2006). corrected (Li et al., 2018; Vermote et al., 2016). During Figure 1. Location of the study area. (a) Asia, (b) Bangladesh, (c) Rema-kalenga Wildlife Sanctuary with plot locations. 4 MOHAMMED ET AL. our analysis, as the path and row of the study area were assess the relationship of these indices with dNBR not specifically indicated, GEE automatically selected index. all the images that intersected with our study area’s In order to determine which month will be appro- boundary – an efficient feature of the GEE platform priate for conducting field-level forest fire investiga- (Gorelick et al., 2017). tion for reference data collection, a time series plot of NBR of the whole study area was made from January 2019 until March 2021. Before creating the Spectral index calculation and analyses time series plot, some preprocessing on the image dataset was done. First of all, the Sentinel-2A dataset For the detection of fire-affected areas, among the fire was filtered using the above-mentioned date range. detecting spectral indices, we used Normalized Burn Additionally, a cloud cover percentage of 10% was Ratio (NBR) and differenced Normalized Burn Ratio used to create an image collection with the least pos- (dNBR) which is calculated using the formula given in sible cloud. Besides, we also used the quality assess- Table 2. In addition to NBR indices, we also calculated, ment (QA60) band of the Sentinel-2A dataset to mask using the formula presented in Table 2, Normalized clouds and cirus clouds in order to create an image vegetation index (NDVI), Differenced Normalized collection with the least possible cloud contamination. vegetation index (dNDVI), Enhanced Vegetation Before calculating NBR indices, B8 bands were Index (EVI), Differenced Enhanced Vegetation Index resampled to 20 m pixel size, similar to the B12 (dEVI), Land Surface Temperature (LST) and bands of the sentinel imageries. Similar resampling Differenced Land Surface Temperature (dLST) to Table 2. Studied indices and formula used to calculate them. Fire Indices Formula Reference B8 B12 NBR (Key & Benson, 2006) B8þB12 dNBR NBR – NBR (Cocke et al., 2005; van Wagtendonk et al., 2004) pre-fire post-fire B8 B4 NDVI (Huete & Jackson, 1987) B8þB4 dNDVI NDVI – NDVI (Jin et al., 2013) Pre-fire Post-fire EVI 2:5ððB8 B4Þ=ðB4þ 6� B4 7:5� B2þ 1ÞÞ (Rocha & Shaver, 2009) dEVI EVI – EVI (Chen et al., 2011) Pre-fire Post-fire Tb 1 LST (Ermida et al., 2020) Ai þ Bi þ Ci ε ε dLST LST – LST (Çolak & Sunar, 2022) Pre-fire Post-fire NBR – Normalized Burn Ratio, dNBR – differenced Normalized Burn Ratio, B2, B4, B8 and B12 refer to the Blue, Red, NIR and SWIR2 bands of the Sentinel-2A dataset. In terms of Land surface temperature (LST) calculation, Tb is the brightness temperature in the TIR band, ε is the surface emissivity for the same band, Ai, Bi and Ci are coefficients. Figure 2. Variation of normalized burn ratio (NBR) values of the studied area from the year 2019 to 2022. GEOLOGY, ECOLOGY, AND LANDSCAPES 5 approach was also done for the calculation of all other selected from the calculated dNBR map layer. From indices (NDVI, EVI and LST). From the time series each of the dNBR classes (Unburned, Burned and plot (Figure 2) it can be seen that NBR value plateaued Regrowth), we selected 10 reference plots (a total of during the months, of March and April, of both the 30 plots) randomly for accuracy assessment of the years 2019 and 2020, indicating the presence of possi- dNBR layer and determined their plot classification ble forest-burned areas during these months. according to the guidelines provided by Key and Therefore, we decided to conduct our field survey Benson (2006) (Figure 3). Each of these plots was of during the month of April 2021. size 20 m × 20 m corresponding to the size of the pixel For our burned area analysis using the NBR indices, of Sentinel datasets. Additionally, we collected soil we incorporated the initial assessment approach (Key & samples and vegetational data for further analysis. Benson, 2006) for burn area analysis where we used a single image from 1 April 2020 (Pre-fire year) – dated Soil collection and laboratory analyses prior to our fieldwork – to calculate the NBR and pre-fire an image from 26 April 2021 (Post-fire year) – dated During soil sample collection from the plot, five after our fieldwork – to calculate the NBR . Both soil samples were taken from each corner and the post-fire of these images contained less than 5% cloud cover. We center of the plot to reduce the heterogeneity. also used the same images and approaches for the Each of the samples was taken at 10 cm depth calculation of NDVI , NDVI , EVI using a soil auger. Then five soil samples from pre-fire Post-fire Pre-fire and EVI indices, and for LST and LST each plot were mixed up well together in the field Post-fire Pre-fire Post- calculation. LST and LST were calcu- and marked as one sample which was used to fire Pre-fire Post-fire lated using the Landsat 8 dataset and using the method measure soil variables. Measurement of soil pH provided by Ermida et al. (2020). Differenced indices, by using pH meter (Hanna HI2211 pH/ORP dNBR, dNDVI, dEVI and dLST, were worked out using meter) following 1:1 slurry pH testing method the formula given in Table 2. After calculating dNBR, it (KALRA, 1995). And Organic Matter percentage was multiplied by a scale factor of 1000 and three of soil was measured using the Loss on Ignition different class was created based on the dNBR value method (Hoogsteen et al., 2018). Next, the Bulk (Table 3). In this study, different ranges of dNBR values density of soil was measured using the core were used according to the guideline of the United method, a direct method for bulk density evalua- States Geological Survey (USGS) (Key & Benson, tion (AL-SHAMMARY et al., 2018). For deter- 2006) and separated them into three classes mining soil texture – sand, silt and clay (Regrowth, Unburned and Burned) (Table 3). The percentages of soil samples – the hydrometer accuracy (users accuracy, producers, overall accuracy method (Faé et al., 2019) was used. A specially and kappa coefficient) of the dNBR index was assessed calibrated hydrometer is used to calculate the using the reference plots collected during the field density of a suspension at a given depth after investigation. Fire hazards are almost non-existent in a period of time setting and a special sedimenta- the context of forest of Bangladesh. In almost all cases, tion cylinder is used to place the soil suspension the source of these fire incidents in the forest of in which the hydrometer is read. Bangladesh is anthropogenic and does not cause wide- spread damage to the overall forest structure similar to Vegetational data collection the forest of other parts of the world. As a result, in this study, these three classes were investigated instead of Apart from soil samples, from every plot, we also separating the classes according to their intensity level collected vegetational data – litter layer (cm), bare as indicated in the guideline provided by Key and ground (%), tree cover (%), shrub cover (%) and Benson (2006). herb cover (%) - to evaluate the degree of relationship of these variable with fire index. Litter layer in the plots were measured using a measurement scale (Marimon-Junior & Hay, 2008). Litter layer measure- Field data collection ments were taken at five random points inside the plot During our field investigation on April 2021, we sur- and average of these reading were calculated. In terms veyed a total of 30 plots, the location of which was of measuring litter depth of burned plots, ash was removed carefully from the top of litter layer in order to only measure the litter. Bare ground (%), Table 3. Severity classes and their respective ranges used in Shrub (%) and Herb cover (%) were visually estimated this study. in each plot, and tree cover (%) was measured using Plot Class dNBR Range a spherical densiometer (McKenzie et al., 2000) at Regrowth − 2000 to −100 Unburned −99 to 100 random five points in the plot and average of these Burned 101 to 2000 five readings were worked out. 6 MOHAMMED ET AL. Figure 3. Field picture of (a), (b) burned plots, (c), (d) unburned plots and (e), (f) regrowth plots. Statistical analysis other variables. We also performed simple linear regression between dNBR and other variables to pre- All of the statistical analyses were performed in the dict the behavior of the studied variables. Additionally, R statistical software version 4.1.1 (R Core Team, a One-way analysis of variances (ANOVA) test was 2017). At first, the descriptive statistics of all the data- performed to assess whether there is any significant sets used in this study were evaluated in order to difference among the plot classes in term of their understand the initial summarization of the data. spectral index values, vegetation and soil properties. Then, the Pearson correlation test was performed, We used the R package “ggstatsplot” to conduct and using the “ggstatsplot” package (Patil, 2021), to inves- visualize the ANOVA test. The significant differences tigate the linear correlation between the dNBR and were tested at P < 0.05. GEOLOGY, ECOLOGY, AND LANDSCAPES 7 Results Soil Spectral indices The summarized information on soil properties is given in Table 4. Mean of values of pH, Organic matter The Descriptive statistics summarize the spectral and Bulk density, Sand, Silt and Clay were 5.02, 1.18, indices, soil properties and vegetation properties 8.15, 68.04, 10.02 and 17.58 respectively with varying information of the field plots (Table 4). Among the ranges. Correlation test, as well as regression test spectral indices, the mean value of dNBR is −13.59 showed that Organic matter (P = 0.000354, r = −0.61, 2 2 with a range between −245.12 and 615.03, whereas the R = 0.3711) and Clay (r = 0.00136, r = −0.56, R = spatial distribution of all the NBR indices is depicted 0.3113) has significant negative relationship with in Figure 4a-c. Additionally, Table 4 also shows the dNBR (Figure 8a, f; Table 7). But, pH (P = 0.01, r = 2 2 mean values of dNDVI (Range: −0.19–0.38), dEVI 0.44, R = 0.1978), Bulk density (P = 0.08, r = 0.32, R = 0.105), Sand (P = 0.01, r = 0.46, R = 0.2111) and Silt (Range: −0.15–0.18) and dLST (Range: −12.79–8.9) (P = 0.0000911, r = 0.65, R = 0.4267) showed positive to be −0.01, −0.01 and −7.05, respectively. The spatial association with dNBR (Figure 8b-e; Table 7). From representation (Figure 4d-i) of EVI and NDVI is the ANOVA result, it can be observed that fire occur- almost similar and dLST map with its general range rence significantly affected all the soil variables value between −0.047 and −13.7 is presented in (Figure 9a-f; Table 7) as almost of the soil property Figure 3i. The classified images of fire severity variables differed significantly (P < 0.05) among the (Figure 5a-c) showed the measured area in terms of plot classes. Burned, Unburned and Regrowth classes (Table 5). The accuracy of the fire severity classification (dNBR) was assessed using the field reference data, Vegetation and the accuracy information is given in Table 6. The According to the descriptive statistics of vegetation total accuracy of the classification was 76.67% and the properties, the mean values of Litter Layer (cm), Bare Kappa coefficient was 0.65. ground (%), Tree cover (%), Shrub cover (%) and Herb Results of the correlation and regression test cover (%) were 8.32, 25.08, 52.31, 21.26 and 18.5, showed that dLST was significantly related (P = 0.01 respectively (Table 4). Among the correlation and and r = −0.44, R = 0.1962) with dNBR which had regression test between different vegetation properties, a negative trend (Figure 6a; Table 7). On the other dNBR had significant negative correlation with Litter hand, dNDVI (P = 0.0000032 and r = −0.74, R = layer (P = 0.0000693, r = −0.66, R = 0.4373), Shrub 0.5452) and dEVI (P = 0.00981 and r = 0.46, R = cover (P = 0.00179, r = 0.55, R = 0.2984) and Herb 0.2152) both indicated positive correlation with fire cover (P = 0.02, r = 0.42, R = 0.1762) (Figure 10a, d, index (Figure 6(b,c); Table 7). We also evaluated the e; Table 7). On the contrary, within the group of significant differences among the Burned, Unburned variables having positive correlation with dNBR, bare and Regrowth plots in terms of the spectral indices ground (P = 0.000722, r = 0.58, R = 0.3399) showed using the One-way ANOVA test. Burned plots were significant association (Figure 10b; Table 7). Similar significantly different from both, Unburned and to soil properties, vegetation properties were signifi- Regrowth plots, in terms of dEVI values, but the cantly affected by fire. In the ANOVA analysis, except dNDVI and dLST showed no statistically noteworthy for Tree cover (P > 0.05), all the variables showed difference (Figure 7a-c; Table 7). Table 4. Descriptive statistics of differenced spectral indices, soil and vegetation properties of all the surveyed plots. Descriptive Statistics Variables Number of sampled plots Minimum Maximum Mean Std. Deviation Spectral indices dLST 30 −12.79 8.9 −7.05 5.54 dNBR 30 −245.12 615.03 −13.59 223.59 dNDVI 30 −0.19 0.388 −0.01 0.14 dEVI 30 −0.15 0.18 −0.02 0.11 Soil pH 30 4 6.1 5.02 0.58 Properties Bulk density (g/cm3) 30 0.81 1.62 1.18 0.23 Organic matter (%) 30 4 13.2 8.15 2.61 Sand (%) 30 50.28 87.5 68.04 10.88 Silt (%) 30 3.58 30 10.02 4.73 Clay (%) 30 3.95 27.78 17.58 6.78 Vegetation Properties Litter Layer (cm) 30 0.9 17 8.32 4.51 Bare ground (%) 30 10 42 25.08 8.15 Tree cover (%) 30 10 74 52.31 19.65 Shrub cover (%) 30 14.5 28 21.26 3.81 Herb cover (%) 30 5 53 18.5 12.57 8 MOHAMMED ET AL. Figure 4. Spectral index maps for the (a) pre-fire NBR, (b) post-fire NBR, (c) dNBR, (d) pre-fire NDVI, (e) post-fire NDVI, (f) dNDVI, (g) pre – fire EVI, (h)post-fire EVI, (i) dEVI, (j) pre-fire LST, (k) post-fire LST, and (l) dLST of the study area. GEOLOGY, ECOLOGY, AND LANDSCAPES 9 Figure 5. Classified map of (a) differenced normalized burn ratio (dNBR), (b) pre-fire normalized burn ratio (Nbr ) and (c) post- pre-fire fire normalized burn ratio (Nbr ) of the studied region divided into 3 classes: unburned, burned and regrowth. post-fire significant differences (P < 0.05) among plot classes 2013; Vivchar, 2011). Therefore, it can be said that (Figure 11a-e; Table 7). forest burning, at present, is a subordinate issue in the study area, but in the future, the scenario might change. The classification accuracy of dNBR was Discussion 76.67% and the Kappa coefficient was 0.65 (Table 6) and these results indicated that the classification of fire Spectral indices severity matches our field data significantly. This study is the first to evaluate forest burn in Several studies reported that dNBR had positive Bangladesh using satellite imageries. Except for LST correlation with dNDVI and dEVI and was a good all of the spectral indices were derived using Sentinel alternative for burn indices in examining burn area datasets – a high-resolution remote sensing dataset – (Dindaroglu et al., 2021; Escuin et al., 2008; Rahman which enables accurate fire severity classification and et al., 2019). Similar results were also found in this analysis of the vegetation indices. The dNDVI and study where both indices, dNDVI and dEVI, showed dEVI showed mean values of −0.01283 and −0.01976 a high correlation with the fire index, meaning that respectively and both of these differenced indices indi- these indices would be effective in explaining fire dis- cated an increase in vegetation cover (Chen et al., turbance. Land surface temperature (LST) is one of the 2011; Jin et al., 2013). Bhuiyan et al. (2021) conducted most important indicators of fire severity as variation a study in Rema-Kalenga Wildlife Sanctuary using in the spatial distribution of LST was observed due to Landsat datasets in which NDVIs of different years fire occurrence in different studies (Montes-Helu were calculated and vegetation growth was noted in et al., 2009; Veraverbeke et al., 2012). (Çolak & successive years. The dNBR index showed the quantity Sunar, 2020) and Veraverbeke et al. (2012) concluded of burned area (Table 5) which is significantly lower that the LST has a significant correlation with fire compared to the magnitude of fire-affected areas in indices. Other studies also reported that after fire other countries (Carreiras et al., 2020; Delcourt et al., LST values of a particular burn site also increase 2021; Dindaroglu et al., 2021; Li et al., 2015; Lindenmayer & Taylor, 2020; Veraverbeke & Hook, (Lambin et al., 2003; Montes-Helu et al., 2009; Table 5. Area of different plot classes according to pre-fire normalized burn ratio (Nbr ), post-fire normalized burn ratio pre-fire (Nbr ) and difference normalized burn ratio (dNBR). post-fire NBR Burned (Sq. km) Unburned (Sq. km) Regrowth (Sq. km) Pre-fire NBR 1.89 60.43 - Post-fire NBR 0.65 61.6788 - dNBR 0.60 59.88 1.84 10 MOHAMMED ET AL. Figure 6. Results of Pearson correlation among differenced normalized burn ratio (dNBR) and different spectral indices: (a) dLST, (b) dNDVI, and (c) dEVI. Veraverbeke et al., 2012; Wendt et al., 2007) which compared to EVI (Yang et al., 2012). This could be results in increasing temperature change in dLST one of the reasons why the NDVI could not differ- index. This study also presented interchangeable entiate the vegetation properties among the plot results meaning that as fire severity (dNBR) exacer- classes (Figure 8b) in this study. Nonetheless, EVI bates, Land surface temperature increases (Figure 6a). can ignore atmospheric and canopy background influ- Rema Kalenga Wildlife Sanctuary has high above- ences (Liu & Huete, 1995) which makes it a more ground biomass (Tipu et al., 2021). In terms of vegeta- reliable index for vegetation monitoring, especially in tion detection, NDVI is more effective than EVI when a high biomass region (Huete et al., 2002); these prop- it comes to a study area with sparse vegetation as erties of EVI help to corroborate our result. dLST NDVI is more sensitive to low biomass conditions values also did not differ significantly across plot Table 6. Results of accuracy assessment of different classes of differenced normalized burn ratio (dNBR) along with overall accuracy and kappa coefficient. Classes Producer’s Accuracy (%) User’s Accuracy (%) Burned 100% 60% Regrowth 100% 70% Unburned 58.90% 100% Overall Accuracy 76.67% Kappa coefficient 0.65 GEOLOGY, ECOLOGY, AND LANDSCAPES 11 Figure 7. Results of one-way ANOVA test among the plot classes for (a) dLST, (b) dNDVI, and (c) dEVI and significant differences were tested at P <0.05. Table 7. Results of accuracy assessment of different classes of differenced normalized burn ratio (dNBR) along with overall accuracy and kappa coefficient. Tests Pearson correlation Simple linear regression ANOVA test Variables r P R P F-value P dLST −0.44 0.01 0.1962 0.01422 0.51 0.61 dNDVI 0.74 0.00000320 0.5452 0.000003197 3.03 0.06 dEVI 0.46 0.00981 0.2152 0.009814 4.96 0.01 OM −0.61 0.000354 0.3711 0.0003536 15.37 0.000035 pH 0.44 0.01 0.1978 0.0138 11.23 0.000283 BD 0.32 0.08 0.105 0.08063 10.05 0.000546 Sand (%) 0.46 0.01 0.2111 0.01064 6.31 0.00564 Silt (%) 0.65 0.0000911 0.4267 0.009109 6.76 0.00419 Clay (%) −0.56 0.00136 0.3113 0.001357 12.33 0.000157 Litter Layer (cm) −0.66 0.0000693 0.4373 0.006932 37.31 0.000000017 Bare Ground (%) 0.58 0.000722 0.3399 0.0007221 10.62 0.000384 Tree Cover (%) 0.20 0.28 0.04146 0.2805 3.42 0.05 Shrub Cover (%) −0.55 0.00179 0.2984 0.001791 11.68 0.000221 Herb Cover (%) −0.42 0.02 0.1762 0.02092 7.88 0.00202 12 MOHAMMED ET AL. Figure 8. Results of Pearson correlation among differenced normalized burn ratio (dNBR) and soil properties: (a) organic matter (%), (b) pH, (c) bulk density (g/cm3), (d) sand (%), (e) silt (%) and (f) clay (%). classes (Figure 8). The anthropogenic fire that ignited 20°C in worst-case scenarios where the average in the study area was not severe as can be concluded increase is 13°C (Vlassova et al., 2014). from the amount of area described in Table 5. Additionally, the statistical summary of dLST Soil properties (Table 4) also indicated a relatively short range of −12.7986°C to 8.9°C. In countries where fire incidents Forest fire is responsible for the changes in the chemi- are more common, temperature increases more than cal, physical and biological properties of forest soil GEOLOGY, ECOLOGY, AND LANDSCAPES 13 Figure 9. Results of one-way ANOVA test among the plot classes for (a) organic matter (%), (b) pH, (c) bulk density (g/cm3), (d) sand (%), (e) silt (%) and (f) clay (%) and significant differences were tested at P <0.05. (Agbeshie et al., 2022) and transferring of heat from matter of soil will increase or decrease. For instance, the fire body in the soil is the main cause of change Badía et al. (2014) reported that due to a high severity (O’Brien et al., 2018). How much of these properties fire event, a significant amount of soil organic matter will change depends upon the intensity, frequency and has diminished. Similarly, Caon et al. (2014) also sug- duration of the fire events (Alcañiz et al., 2018; gested that organic matter is negatively associated with Fernández-García et al., 2021; Lucas borja et al., fire severity which is similar to the finding of this study 2021). Depending on the fire intensity, duration, avail- (Figure 8a). Also, in this study, clay showed a negative able biomass and fire type (Reyes et al., 2015), organic 14 MOHAMMED ET AL. Figure 10. Results of Pearson correlation between differenced normalized burn ratio (dnbr) and vegetation properties: (a) litter layer (cm), (b) bare ground (%), (c) tree cover (%), (d) shrub cover (%) and (e) herb cover (%). relationship with fire index (Figure 8f). Compared to the fact that both sand and silt have a positive and clay sand and silt, clay particles of soil are more easily negative correlation with dNBR in this study. With affected by fire as the temperature threshold of clay increasing fire severity, pH value in our study area also is lower than sand and silt (Alcañiz et al., 2016; Neary increases (Figure 8b). Generally, with elevated fire et al., 2005). Higher forest fire intensity leads to the severity, pH value of soil also increases because of collapse of clay particles and produces more silt and the addition of nutrient-rich ash content – produced sand in the soil (Alcañiz et al., 2018) which explains due to the burning of organic matter in soil (Alcañiz GEOLOGY, ECOLOGY, AND LANDSCAPES 15 Figure 11. Results of one-way ANOVA test among the plot classes for (a) litter layer (cm), (b) bare ground (%), (c) tree cover (%), (d) shrub cover (%) and (e) herb cover (%) and significant differences were tested at P <0.05. et al., 2018). Several studies also reported an increase et al., 2021; Fultz et al., 2016). The bulk density result in soil pH after forest fire of different types (Alcañiz of this study showed no significant correlation with et al., 2016; Dzwonko et al., 2015; Francos et al., 2019). fire severity (Figure 8c). Fire has mixed results on the But, some studies also indicated that pH level did not particular properties of soil; some studies showed change after occurrence of fire (Badía et al., 2014; a positive correlation of fire with bulk density Fernández-Fernández et al., 2015; Fernández-García (Granged et al., 2011; Heydari et al., 2017; Jordán 16 MOHAMMED ET AL. et al., 2011) while others observed the opposite (Chief study indicated bare ground has a positive association with fire severity (Figure 10b). (de Benavides-Solorio et al., 2012; Downing et al., 2017). & MacDonald, 2005) also reported that bare ground Among all the soil properties, only organic matter was a good indicator of forest fire severity and fire the showed significant differences among burned, unburned bare soil amount went higher as the fire severity ele- and recovered areas (Figure 9a). However, all the other vated. Regarding tree cover, which is not significantly properties differed significantly between burned and related to fire index, several plots in this study suffered unburned zone (Figure 9b-f), therefore suggesting that the burning of vegetation in the understory layer and the studied fire had severe consequences on the soil the forest floor, denoting the burning of only shrub properties. Organic matter recovery in soil affected by and herb cover in the forest plots and the fire did not affect the tree cover as it not severe. As a result, shrub fire depends on the microbial activity in the soil (Muñoz- and herb cover showed a significant negative correla- Rojas et al., 2016). Availability of nutrients and increased tion with fire intensity. Similar findings were also soil moisture and pH after fire can help to elevate the observed by other studies studying the relationship microbial activities in the soil which in turn can contri- between field data and spectral indices (Cardil et al., bute to organic matter increase (MABUHAY et al., 2003; 2019; Miller & Quayle, 2015; Miller et al., 2009). Muñoz-Rojas et al., 2016). In this study, the organic Results of the ANOVA test in this study sug- matter values in the regrowth zones were significantly gested all the vegetation properties, other than tree different from both burned and unburned classes, mean- cover, were significantly affected by fire ing fire disturbed soil from the pre-fire year started its (Figure 11a-e). As stated previously in the correla- recovery according to the rationales stated previously. tion analysis of tree cover that the tree cover was not affected by fire like other properties, the similar Regarding Clay, recovery starts as soon as the distur- conclusion can also be drawn from the results of bance ceases and the burned portion of the soil starts ANOVA analysis. Shrub cover in our study area recovering (Chen et al., 2009; Zhao et al., 2010, 2011). But was significantly higher in regrowth plots than in the recovery of clay usually takes a long period of time the burned plots, but the herb cover did not show (Zhao et al., 2010) which is illustrated in our study as the similar pattern. This showed that shrub cover was clay particles in the regrowth area were significantly recovering faster than herb. But, this result com- lower than in the unburned area as clay particles in the pletely contradicts the studies conducted by (Calvo disturbed plots from the pre-fire year still have not fully et al., 2005; Vega et al., 2014). As soon as the recovered. Decrease in soil pH after fire usually takes regeneration of the burned area starts to occur from 3 months up to several years (Granged et al., 2011; the bare soil (%) starts to decrease (de Benavides- Jordán et al., 2010) and is initiated by the removal of fire Solorio & MacDonald, 2005; Meyer et al., 2004). ash from the soil. In this study, pH value of regrowth plots and of unburned plots showed no significant dif- Conclusion ference, therefore it can be concluded that pre-fire dis- turbance is slowly subsidizing. Bulk density result in our The results of this study helped to understand the study also shows a similar pattern of recovery. association of fire index with other spectral indices, soil and vegetation properties, and showed how the spectral indices, soil and vegetation properties differed Vegetation properties significantly among the three plots classes (burned, unburned and regrowth), indicating the significant Correlation results of the vegetation properties showed that except for Tree cover (%) all the other effect of fire, as well as regrowth dynamics of burned properties were significantly correlated (Figure 10a-e). vegetation from the previous year. Besides, from the Litter layer depth changes generally depend on the fire relationship of the burn index with other spectral severity level, combustion, heat transfer, and amount indices, it is evident that other indices can also detect of litter layer present on the soil surface (Neary et al., the fire-affected area. Even though the measured area 2005). Usually, after fire, a thick ash layer would be of the burned forest was not significant, the knowledge created on the litter layer due to charring. Further and information of this study will help the local autho- investigation of the pre-fire litter layer is required as rities to be aware of the combustion potential of this it will indicate the litter depth to ash depth ratio which forest ecosystem and take necessary measures in the will, in turn, suggest fire severity in this area. Ash conservation management, keeping in mind the global depth was not measured in this study as the study climate change scenarios, to alleviate future fire poten- area is wind prevalent (NSP, 2006) and might cause redistribution of ash layer throughout the area. Our tial as this is a novel study involving forest fire using GEOLOGY, ECOLOGY, AND LANDSCAPES 17 remote sensing techniques in Bangladesh. Further stu- Changes at cm-scale topsoil. CATENA, 113, 267–275. https://doi.org/10.1016/j.catena.2013.08.002 dies should be conducted using more spectral indices Bhuiyan, M. S., Islam, S., Haque, M. M. U., Aktar, S., & and vegetation and soil parameters to evaluate severity Ahmed, R. (2021). Evaluating capital assets in governing relation. Additionally, we can also use more complex protected area co-management in the Rema-Kalenga statistical measures to derive accurate relationships Wildlife Sanctuary, Bangladesh. International Forestry Review , 23 (1), 16–28. https://doi.org/10.1505/ from these variables. Bowman, D. M. J. S., Kolden, C. A., Abatzoglou, J. T., Johnston, F. H., van der Werf, G. R., & Flannigan, M. Acknowledgments (2020). Vegetation fires in the anthropocene. Nature We are grateful to the faculty members of Forestry and Reviews Earth & Environment, 1(10), 500–515. https:// Environmental Sciences, Shahjalal University of Science doi.org/10.1038/s43017-020-0085-3 and Technology, Bangladesh, for their immense support. Calvo, L., Tárrega, R., Luis, E., Valbuena, L., & Marcos, E. Also, we would like to express our gratitude to all of the (2005). Recovery after experimental cutting and burning people who provided support during the research work. in three shrub communities with different dominant species. Plant Ecology, 180(2), 175–185. https://doi.org/ 10.1007/s11258-005-0200-z Disclosure statement Caon, L., Vallejo, V. R., Ritsema, C. J., & Geissen, V. (2014). Effects of wildfire on soil nutrients in Mediterranean The authors have no relevant financial or non-financial ecosystems. Earth-Science Reviews, 139, 47–58. https:// interests to disclose. doi.org/10.1016/j.earscirev.2014.09.001 Cardil, A., Mola-Yudego, B., Blázquez-Casado, Á., & González-Olabarria, J. R. (2019). Fire and burn severity ORCID assessment: Calibration oF relative differenced normal- ized burn ratio (RdNBR) with field data. Journal of Mohammed http://orcid.org/0000-0002-7614-0013 Environmental Management, 235, 342–349. https://doi. Ariful Khan http://orcid.org/0000-0001-8169-1277 org/10.1016/j.jenvman.2019.01.077 Angana Kuri http://orcid.org/0000-0002-7381-6188 CARDOSO, M. F., HURTT, G. C., MOORE, B., Sohag Ahammed http://orcid.org/0000-0002-9162-178X NOBRE, C. A., & PRINS, E. M. (2003). Projecting future Kazi Al Muqtadir Abir http://orcid.org/0000-0003-2314- fire activity in Amazonia. Global Change Biology, 9(5), 656–669. https://doi.org/10.1046/j.1365-2486.2003.00607.x Mohammed A.S. Arfin-Khan http://orcid.org/0000- Carreiras, J. M. B., Quegan, S., Tansey, K., & Page, S. (2020). 0001-6275-7023 Sentinel-1 observation frequency significantly increases burnt area detectability in tropical SE Asia. Environmental Research Letters, 15(5), 054008. https://doi.org/10.1088/ References 1748-9326/ab7765 Agbeshie, A. A., Abugre, S., Atta Darkwa, T., & Awuah, R. Catry, F. X., Rego, F. C., Bação, F. L., & Moreira, F. (2009). (2022). A review of the effects of forest fire on soil Modeling and mapping wildfire ignition risk in Portugal. properties. Journal of Forestry Research, 33(5), International Journal of Wildland Fire, 18(8), 921. https:// 1419–1441. https://doi.org/10.1007/s11676-022-01475-4 doi.org/10.1071/WF07123 Alcañiz, M., Outeiro, L., Francos, M., Farguell, J., & Chen, X., Vogelmann, J. E., Rollins, M., Ohlen, D., Key, C. H., Úbeda, X. (2016). Long-term dynamics of soil chemical Yang, L., Huang, C., & Shi, H. (2011). Detecting post-fire properties after a prescribed fire in a Mediterranean for- burn severity and vegetation recovery using multitemporal remote sensing spectral indices and field-collected compo- est (Montgrí Massif, Catalonia, Spain). The Science of the site burn index data in a ponderosa pine forest. Total Environment, 572, 1329–1335. https://doi.org/10. International Journal of Remote Sensing, 32(23), 1016/j.scitotenv.2016.01.115 7905–7927. https://doi.org/10.1080/01431161.2010.524678 Alcañiz, M., Outeiro, L., Francos, M., & Úbeda, X. (2018). Chen, R., Zhang, Y., Li, Y., Wei, W., Zhang, J., & Wu, N. Effects of prescribed fires on soil properties: A review. (2009). The variation of morphological features and The Science of the Total Environment, 613–614, 944–957. mineralogical components of biological soil crusts in the https://doi.org/10.1016/j.scitotenv.2017.09.144 Gurbantunggut Desert of Northwestern China. AL-SHAMMARY, A. A. G., KOUZANI, A. Z., KAYNAK, A., Environmental Geology, 57(5), 1135–1143. https://doi. KHOO, S. Y., NORTON, M., & GATES, W. (2018). Soil org/10.1007/s00254-008-1410-1 bulk density estimation methods: A review. Pedosphere, 28 Chief, K., Young, M. H., & Shafer, D. S. (2012). Changes in (4), 581–596. https://doi.org/10.1016/S1002-0160(18) soil structure and hydraulic properties in a wooded 60034-7 shrubland ecosystem following a prescribed fire. Soil Arndt, N., Vacik, H., Koch, V., Arpaci, A., & Gossow, H. Science Society of America Journal, 76(6), 1965–1977. (2013). Modeling human-caused forest fire ignition for https://doi.org/10.2136/sssaj2011.0072 assessing forest fire danger in Austria. IForest - Chowdhury, M. S. H., & Koike, M. (2010). Therapeutic use Biogeosciences and Forestry, 6(6), 315–325. https://doi. of plants by local communities in and around org/10.3832/ifor0936-006 Rema-Kalenga Wildlife Sanctuary: Implications for pro- Badía, D., Martí, C., Aguirre, A. J., Aznar, J. M., González- tected area management in Bangladesh. Agroforestry Pérez, J. A., De la Rosa, J. M., León, J., Ibarra, P., & Echeverría, T. (2014). Wildfire effects on nutrients and Systems, 80(2), 241–257. https://doi.org/10.1007/s10457- organic carbon of a Rendzic Phaeozem in NE Spain: 010-9316-9 18 MOHAMMED ET AL. Chowdhury, M. R., & Ndiaye, O. (2017). Climate change of Remote Sensing, 29(4), 1053–1073. https://doi.org/10. and variability impacts on the forests of Bangladesh - 1080/01431160701281072 a diagnostic discussion based on CMIP5 GCMs and Faé, G. S., Montes, F., Bazilevskaya, E., Añó, R. M., & ENSO. International Journal of Climatology, 37(14), Kemanian, A. R. (2019). Making soil particle size analysis 4768–4782. https://doi.org/10.1002/joc.5120 by laser diffraction compatible with standard soil texture Chuvieco, E., Aguado, I., Yebra, M., Nieto, H., Salas, J., determination methods. Soil Science Society of America Martín, M. P., Vilar, L., Martínez, J., Martín, S., Journal, 83(4), 1244–1252. https://doi.org/10.2136/ Ibarra, P., de la Riva, J., Baeza, J., Rodríguez, F., sssaj2018.10.0385 Molina, J. R., Herrera, M. A., & Zamora, R. (2010). Fernández-Fernández, M., Gómez-Rey, M. X., & González- Development of a framework for fire risk assessment Prieto, S. J. (2015). Effects of fire and three fire-fighting using remote sensing and geographic information system chemicals on main soil properties, plant nutrient content technologies. Ecological modelling, 221(1), 46–58. https:// and vegetation growth and cover after 10 years. The doi.org/10.1016/j.ecolmodel.2008.11.017 Science of the Total Environment, 515–516, 92–100. Cocke, A. E., Fulé, P. Z., & Crouse, J. E. (2005). Comparison https://doi.org/10.1016/j.scitotenv.2015.02.048 of burn severity assessments using differenced normal- Fernández-García, V., Marcos, E., Huerta, S., & Calvo, L. ized burn ratio and ground data. International Journal of (2021). Soil-vegetation relationships in Mediterranean Wildland Fire, 14(2), 189. https://doi.org/10.1071/ forests after fire. Forest Ecosystems, 8(1), 18. https://doi. WF04010 org/10.1186/s40663-021-00295-y Çolak, E., & Sunar, F. (2020). Spatial pattern analysis of Flannigan, M. D., Amiro, B. D., Logan, K. A., Stocks, B. J., & post-fire damages in the menderes district of Turkey. Wotton, B. M. (2006). Forest fires and climate change in Frontiers of Earth Science, 14(2), 446–461. https://doi. the 21st century. Mitigation and Adaptation Strategies for org/10.1007/s11707-019-0786-4 Global Change, 11(4), 847–859. https://doi.org/10.1007/ Çolak, E., & Sunar, F. (2022). Investigating the usefulness of s11027-005-9020-7 satellite-retrieved land surface temperature (LST) in pre- Flannigan, M., Cantin, A. S., de Groot, W. J., Wotton, M., Newbery, A., & Gowman, L. M. (2013). Global wildland and post-fire spatial analysis. Earth Science Informatics. fire season severity in the 21st century. Forest Ecology and https://doi.org/10.1007/s12145-022-00883-8 Management, 294, 54–61. https://doi.org/10.1016/j. de Benavides-Solorio, J. D., & MacDonald, L. H. (2005). foreco.2012.10.022 Measurement and prediction of post-fire erosion at the Francos, M., Stefanuto, E. B., Úbeda, X., & Pereira, P. hillslope scale, Colorado front range. International Journal of Wildland Fire, 14(4), 457. https://doi.org/10. (2019). Long-term impact of prescribed fire on soil 1071/WF05042 chemical properties in a wildland-urban interface. Delcourt, C. J. F., Combee, A., Izbicki, B., Mack, M. C., Northeastern Iberian Peninsula. The Science of the Maximov, T., Petrov, R., Rogers, B. M., Scholten, R. C., Total Environment, 689, 305–311. https://doi.org/10. Shestakova, T. A., van Wees, D., & Veraverbeke, S. 1016/j.scitotenv.2019.06.434 (2021). Evaluating The differenced normalized burn Fultz, L. M., Moore-Kucera, J., Dathe, J., Davinic, M., ratio for assessing fire severity using sentinel-2 imagery Perry, G., Wester, D., Schwilk, D. W., & Rideout- in Northeast Siberian larch forests. Remote Sensing, 13 Hanzak, S. (2016). Forest wildfire and grassland pre- (12), 2311. https://doi.org/10.3390/rs13122311 scribed fire effects on soil biogeochemical processes and De Santis, A., & Chuvieco, E. (2007). Burn severity estima- microbial communities: Two case studies in the semi-arid tion from remotely sensed data: Performance of simula- Southwest. Applied Soil Ecology, 99, 118–128. https://doi. tion versus empirical models. Remote sensing of org/10.1016/j.apsoil.2015.10.023 environment, 108(4), 422–435. https://doi.org/10.1016/j. Ganteaume, A., Camia, A., Jappiot, M., San-Miguel- rse.2006.11.022 Ayanz, J., Long-Fournel, M., & Lampin, C. (2013). Dindaroglu, T., Babur, E., Yakupoglu, T., Rodrigo- A review of the main driving factors of forest fire ignition Comino, J., & Cerdà, A. (2021). Evaluation of geomor- over Europe. Environmental management, 51(3), phometric characteristics and soil properties after 651–662. https://doi.org/10.1007/s00267-012-9961-z a wildfire using sentinel-2 MSI imagery for future Gorelick, N., Hancher, M., Dixon, M., Ilyushchenko, S., fire-safe forest. Fire safety journal, 122, 103318. https:// Thau, D., & Moore, R. (2017). Google earth engine: doi.org/10.1016/j.firesaf.2021.103318 Planetary-scale geospatial analysis for everyone. Remote Downing, T. A., Imo, M., Kimanzi, J., & Otinga, A. N. Sensing of Environment, 202, 18–27. https://doi.org/10. (2017). Effects of wildland fire on the tropical alpine 1016/j.rse.2017.06.031 moorlands of Mount Kenya. CATENA, 149, 300–308. Granged, A. J. P., Zavala, L. M., Jordán, A., & Bárcenas- https://doi.org/10.1016/j.catena.2016.10.003 Dzwonko, Z., Loster, S., & Gawroński, S. (2015). Impact of Moreno, G. (2011). Post-fire evolution of soil properties fire severity on soil properties and the development of and vegetation cover in a Mediterranean heathland after tree and shrub species in a Scots pine moist forest site in experimental burning: A 3-year study. Geoderma, 164(1– southern Poland. Forest Ecology and Management, 342, 2), 85–94. https://doi.org/10.1016/j.geoderma.2011.05.017 56–63. https://doi.org/10.1016/j.foreco.2015.01.013 Halim, M. A., Shahid, A., Chowdhury, M. S. H., Ermida, S. L., Soares, P., Mantas, V., Göttsche, F. -M., & Nahar, M. N., Sohel, M. S. I., Nuruddin, M. J., & Trigo, I. F. (2020). Google earth engine open-source code Koike, M. (2008). Evaluation of land-use pattern change for land surface temperature estimation from the Landsat in West Bhanugach Reserved Forest, Bangladesh, using series. Remote Sensing, 12(9), 1471. https://doi.org/10. remote sensing and GIS techniques. Journal of Forestry 3390/rs12091471 Research, 19(3), 193–198. https://doi.org/10.1007/s11676- Escuin, S., Navarro, R., & Fernández, P. (2008). Fire severity assessment by using NBR (normalized burn ratio) and 008-0044-1 NDVi (normalized difference vegetation index) derived Hao, B., Xu, X., Wu, F., & Tan, L. (2022). Long-term effects from LANDSAT TM/ETM images. International Journal of fire severity and climatic factors on post-forest-fire GEOLOGY, ECOLOGY, AND LANDSCAPES 19 vegetation recovery. Forests, 13(6), 883. https://doi.org/ Key, C. H., & Benson, N. C. (2006). Landscape Assessment: 10.3390/f13060883 Ground Measure of Severity, the Composite Burn Index; Hasan, M. E., Zhang, L., Mahmood, R., Guo, H., & Li, G. and Remote Sensing of Severity,the Normalized Burn (2021). Modeling of forest ecosystem degradation due to Ratio. In FIREMON: Fire effects monitoring and inventory anthropogenic stress: The case of Rohingya influx into system (Vol. Gen. Tech. Rep. RMRS-GTR-164-CD, pp. the cox’s bazar–teknaf peninsula of bangladesh. LA1–51). U.S. Department of Agriculture, Forest Service, Environments - MDPI, 8(11). https://doi.org/10.3390/ Rocky Mountain Research Station. environments8110121 Khan, M. S., Hassan, M. A., & Uddin, M. Z. (2002). Hasnat, G. N. T., Kabir, M. A., & Hossain, M. A. (2018). Ethnobotanical survey in Rema-Kalenga Wildlife Major environmental issues and problems of South Sanctuary (Habiganj) in Bangladesh. Bangladesh Journal Asia, particularly Bangladesh. In Handbook of environ- Plant Taxon, 9(1), 51–60. mental materials management (pp. 1–40). Springer Korovin, G. N.1996. Analysis of the distribution of forest fires in International Publishing. https://doi.org/10.1007/978- Russia Bt - fire in ecosystems of Boreal Eurasia. 3-319-58538-3_7-1 J. G. Goldammer & V. V. Furyaev Eds. Springer Heydari, M., Rostamy, A., Najafi, F., & Dey, D. C. (2017). Netherlands. https://doi.org/10.1007/978-94-015-8737-2_8 Effect of fire severity on physical and biochemical soil Lambin, E. F., Goyvaerts, K., & Petit, C. (2003). Remotely- properties in Zagros oak (Quercus brantii Lindl.) forests sensed indicators of burning efficiency of savannah and in Iran. Journal of Forestry Research, 28(1), 95–104. forest fires. International Journal of Remote Sensing, 24 https://doi.org/10.1007/s11676-016-0299-x (15), 3105–3118. https://doi.org/10.1080/ Hoogsteen, M. J. J., Lantinga, E. A., Bakker, E. J., & Tittonell, P. A. (2018). An evaluation of the loss-on-igni- Li, Y., Chen, J., Ma, Q., Zhang, H. K., & Liu, J. (2018). tion method for determining the soil organic matter Evaluation of sentinel-2A surface reflectance derived content of calcareous soils. Communications in Soil using sen2Cor in North America. IEEE Journal of Science and Plant Analysis, 49(13), 1541–1552. https:// Selected Topics in Applied Earth Observations and doi.org/10.1080/00103624.2018.1474475 Remote Sensing, 11(6), 1997–2021. https://doi.org/10. Huete, A., Didan, K., Miura, T., Rodriguez, E., Gao, X., & 1109/JSTARS.2018.2835823 Ferreira, L. (2002). Overview of the radiometric and bio- Lindenmayer, D., & Taylor, C. (2020). Extensive recent physical performance of the MODIS vegetation indices. wildfires demand more stringent protection of critical Remote Sensing of Environment, 83(1–2), 195–213. old growth forest. Pacific Conservation Biology, 26(4), https://doi.org/10.1016/S0034-4257(02)00096-2 384. https://doi.org/10.1071/PC20037 Huete, A. R., & Jackson, R. D. (1987). Suitability of spectral Li, J., Song, Y., Huang, X., & Li, M. (2015). Comparison of indices for evaluating vegetation characteristics on arid rangelands. Remote Sensing of Environment, 23(2), 213– forest burned areas in mainland China derived from IN8. https://doi.org/10.1016/0034-4257(87)90038-1 MCD45A1 and data recorded in yearbooks from 2001 Humayun-Bin-Akram, M., & Masum, K. M. (2020). Forest to 2011. International Journal of Wildland Fire, 24(1), degradation assessment of ratargul special biodiversity 103. https://doi.org/10.1071/WF14031 protection area for conservation implications. Forestist, Liu, H. Q., & Huete, A. (1995). A feedback based modifica- 70(2), 77–84. https://doi.org/10.5152/forestist.2020.20016 tion of the NDVI to minimize canopy background and Islam, M., Deb, G. P., & Rahman, M. (2017). Forest frag- atmospheric noise. IEEE Transactions on Geoscience and mentation reduced carbon storage in a moist tropical Remote Sensing, 33(2), 457–465. https://doi.org/10.1109/ forest in Bangladesh: Implications for policy TGRS.1995.8746027 development. Land Use Policy, 65, 15–25. https://doi. Lucas-borja, M. E., Delgado baquerizo, M., Muñoz org/10.1016/j.landusepol.2017.03.025 rojas, M., Plaza álvarez, P. A., Gómez Sanchez, M. E., Jin, S., Yang, L., Danielson, P., Homer, C., Fry, J., & Xian, G. González romero, J., Peña molina, E., Moya, D., & de (2013). A comprehensive change detection method for las Heras, J. (2021). Changes in ecosystem properties after updating the national land cover database to circa 2011. post fire management strategies in wildfire affected Remote sensing of environment, 132, 159–175. https://doi. Mediterranean forests. The Journal of Applied Ecology, org/10.1016/j.rse.2013.01.012 58(4), 836–846. https://doi.org/10.1111/1365-2664.13819 Jolly, W. M., Cochrane, M. A., Freeborn, P. H., Holden, Z. A., MABUHAY, J. A., NAKAGOSHI, N., & HORIKOSHI, T. Brown, T. J., Williamson, G. J., & Bowman, D. M. J. S. (2003). Microbial biomass and abundance after forest fire (2015). Climate-induced variations in global wildfire dan- in pine forests in Japan. Ecological Research, 18(4), ger from 1979 to 2013. Nature Communications, 6(1), 431–441. https://doi.org/10.1046/j.1440-1703.2003.00567.x 7537. https://doi.org/10.1038/ncomms8537 Mallinis, G., Mitsopoulos, I., & Chrysafi, I. (2018). Jordán, A., González, F. A., & Zavala, L. M. (2010). Re- Evaluating and comparing sentinel 2A and Landsat-8 establishment of soil water repellency after destruction operational land imager (OLI) spectral indices for esti- by intense burning in a Mediterranean heathland (SW mating fire severity in a Mediterranean pine ecosystem of Spain). Hydrological processes, 24(6), 736–748. https:// Greece. GIScience & Remote Sensing, 55(1), 1–18. https:// doi.org/10.1002/hyp.7519 doi.org/10.1080/15481603.2017.1354803 Jordán, A., Zavala, L. M., Mataix-Solera, J., Nava, A. L., & Marimon-Junior, B. H., & Hay, J. D. (2008). A new instru- Alanís, N. (2011). Effect of fire severity on water repel- ment for measurement and collection of quantitative lency and aggregate stability on Mexican volcanic soils. samples of the litter layer in forests. Forest Ecology and CATENA, 84(3), 136–147. https://doi.org/10.1016/j. Management, 255(7), 2244–2250. https://doi.org/10. catena.2010.10.007 1016/j.foreco.2008.01.037 KALRA, Y. (1995). Determination of pH of soils by different McKenzie, D., Halpern, C. B., & Nelson, C. R. (2000). methods: Collaborative study. Journal of AOAC Overstory influences on herb and shrub communities in International, 78(2), 310. https://doi.org/10.1093/jaoac/ 78.2.310 mature forests of western Washington, U.S.A. Canadian 20 MOHAMMED ET AL. Journal of Forest Research, 30(10), 1655–1666. https://doi. Environment, 44(11), 3471–3489. https://doi.org/10. 1111/pce.14176 org/10.1139/x00-091 Meyer, V. F., Redente, E. F., Barbarick, K. A., Brobst, R. B., NSP. (2006) . Nishorgo support project. Management Plans for Rema-Kalenga Wildlife Sanctuary. Paschke, M. W., & Miller, A. L. (2004). Plant and soil O’Brien, J. J., Hiers, J. K., Varner, J. M., Hoffman, C. M., responses to biosolids application following forest fire. Dickinson, M. B., Michaletz, S. T., Loudermilk, E. L., & Journal of Environment Quality, 33(3), 873. https://doi. Butler, B. W. (2018). Advances in mechanistic org/10.2134/jeq2004.0873 Miller, J. D., Knapp, E. E., Key, C. H., Skinner, C. N., approaches to quantifying biophysical fire effects. Current Forestry Reports, 4(4), 161–177. https://doi.org/ Isbell, C. J., Creasy, R. M., & Sherlock, J. W. (2009). 10.1007/s40725-018-0082-7 Calibration and validation of the relative differenced nor- Ommi, A., Emami, F., Zíková, N., Hopke, P. K., & malized burn ratio (RdNBR) to three measures of fire severity in the Sierra Nevada and Klamath Mountains, Begum, B. A. (2017). Trajectory-based models and California, USA. Remote Sensing of Environment, 113(3), remote sensing for biomass burning assessment in Bangladesh. Aerosol and Air Quality Research, 17(2), 645–656. https://doi.org/10.1016/j.rse.2008.11.009 465–475. https://doi.org/10.4209/aaqr.2016.07.0304 Miller, J. D., & Quayle, B. (2015). Calibration and validation Parks, S. A., Holsinger, L. M., Panunto, M. H., Jolly, W. M., of immediate post-fire satellite-derived data to three severity metrics. Fire Ecology, 11(2), 12–30. https://doi. Dobrowski, S. Z., & Dillon, G. K. (2018). High-severity org/10.4996/fireecology.1102012 fire: Evaluating its key drivers and mapping its probabil- ity across western US forests. Environmental Research Mohammed Bhuiyan, A. R., Khan, A., Mia, A., Sarker, A., & Letters, 13(4), 044037. https://doi.org/10.1088/1748- Khan, A. (2022). Effect of transportation infrastructure 9326/aab791 on forest plant diversity and soil properties in Lawachara National Park, Bangladesh. Acta Ecologica Sinica, 42(1), Patil, I. (2021). Visualizations with statistical details: The 110–120. https://doi.org/10.1016/j.chnaes.2021.08.015 “ggstatsplot” approach. Journal of Open Source Software, 6(61), 3167. https://doi.org/10.21105/joss.03167 Mollicone, D., Eva, H. D., & Achard, F. (2006). Human role Rahman, S., Chang, H. -C., Magill, C., Tomkins, K., & in Russian wild fires. Nature, 440(7083), 436–437. https:// Hehir, W. (2019). Spatio-temporal assessment of fire doi.org/10.1038/440436a Montes-Helu, M. C., Kolb, T., Dore, S., Sullivan, B., severity and vegetation recovery utilising sentinel-2 ima- Hart, S. C., Koch, G., & Hungate, B. A. (2009). Persistent gery in New South Wales, Australia. IGARSS 2019 - 2019 IEEE International Geoscience and Remote Sensing effects of fire-induced vegetation change on energy parti- Symposium, 9960–9963. https://doi.org/10.1109/ tioning and evapotranspiration in ponderosa pine forests. IGARSS.2019.8899242 Agricultural and Forest Meteorology, 149(3–4), 491–500. https://doi.org/10.1016/j.agrformet.2008.09.011 Rahman, M. M., Rahman, M. M., Guogang, Z., & Moody, J. A., Ebel, B. A., Nyman, P., Martin, D. A., Stoof, C., & Islam, K. S. (2010). A review of the present threats to tropical moist deciduous Sal (Shorea Robusta) forest eco- McKinley, R. (2016). Relations between soil hydraulic prop- system of Central Bangladesh. Tropical Conservation erties and burn severity. International Journal of Wildland Fire, 25(3), 279. https://doi.org/10.1071/WF14062 Science , 3 (1), 90–102. https://doi.org/10.1177/ Mouillot, F., & Field, C. B. (2005). Fire history and the global 194008291000300108 carbon budget: A 1ox 1o fire history reconstruction for Rana, S., & Hafizur Rahman, S. (2021). Forest cover change analysis using remote sensing and its impact on forest the 20th century. Global Change Biology, 11(3), 398–420. sustainability at Fasiakhali Wildlife Sanctuary., https://doi.org/10.1111/j.1365-2486.2005.00920.x Mukul, S. A., Sohel, M. S. I., Herbohn, J., Inostroza, L., & International Electronic Conference on Forests — König, H. (2017). Integrating ecosystem services supply Sustainable Forests: Ecology, Management, Products potential from future land-use scenarios in protected area and Trade, 1–15 September 2021, Switzerland, 2021 management: A Bangladesh case study. Ecosystem Services, (MDPI: Basel), doi:10.3390/IECF2021-10818 R Core Team. (2017). R: A Language and Environment for 26, 355–364. https://doi.org/10.1016/j.ecoser.2017.04.001 Muñoz-Rojas, M., Erickson, T. E., Martini, D., Dixon, K. W., Statistical Computing. https://www.r-project.org/ & Merritt, D. J. (2016). Soil physicochemical and micro- Reddy, C. S., Bird, N. G., Sreelakshmi, S., Manikandan, T. M., biological indicators of short, medium and long term Asra, M., Krishna, P. H., Jha, C. S., Rao, P. V. N., & post-fire recovery in semi-arid ecosystems. Ecological Diwakar, P. G. (2019). Identification and characterization of spatio-temporal hotspots of forest fires in South Asia. indicators, 63, 14–22. https://doi.org/10.1016/j.ecolind. 2015.11.038 Environmental Monitoring and Assessment, 191(S3), 791. Neary, D. G., Ryan, K. C., & DeBano, L. F. (2005). Wildland https://doi.org/10.1007/s10661-019-7695-6 fire in ecosystems: effects of fire on soils and water. Gen. Redowan, M., Phinn, S., Roelfsema, C., & Aziz, A. A. (2019). Tech. Rep. RMRS-GTR-42-vol.4. Ogden. U.S. REDD+ project design study for quantifying activity data Department of Agriculture, Forest Service, Rocky for historic forest degradation in a Bangladesh forest Mountain Research Station. https://doi.org/10.2737/ using Landsat data. Journal of Applied Remote Sensing, RMRS-GTR-42-V4 13(04), 1. https://doi.org/10.1117/1.jrs.13.046518 Niklasson, M., & Granström, A. (2000). Numbers and sizes Redowan, M., Phinn, S., Roelfsema, C., & Aziz, A. A. (2020). of fires: Long-term spatially explicit fire history in Claslite unmixing of Landsat images to estimate REDD+ a Swedish boreal landscape. Ecology, 81(6), 1484–1499. activity data for deforestation in a Bangladesh forest. https://doi.org/10.1890/0012-9658(2000)081[1484: Journal of Applied Remote Sensing, 14(02), 1. https://doi. NASOFL]2.0.CO;2 org/10.1117/1.jrs.14.024505 Nolan, R. H., Collins, L., Leigh, A., Ooi, M. K. J., Reineking, B., Weibel, P., Conedera, M., & Bugmann, H. Curran, T. J., Fairman, T. A., Resco de Dios, V., & (2010). Environmental determinants of lightning- v. Bradstock, R. (2021). Limits to post fire vegetation human-induced forest fire ignitions differ in a temperate recovery under climate change. Plant, Cell & mountain region of Switzerland. International Journal of GEOLOGY, ECOLOGY, AND LANDSCAPES 21 Wildland Fire, 19(5), 541. https://doi.org/10.1071/ van Wagtendonk, J. W., Root, R. R., & Key, C. H. (2004). WF08206 Comparison of AVIRIS and Landsat ETM+ detection Reyes, O., García-Duro, J., & Salgado, J. (2015). Fire affects capabilities for burn severity. Remote Sensing of soil organic matter and the emergence of Pinus radiata Environment, 92(3), 397–408. https://doi.org/10.1016/j. seedlings. Annals of Forest Science, 72(2), 267–275. rse.2003.12.015 https://doi.org/10.1007/s13595-014-0427-8 Vega, J. A., Fernández, C., Fonturbel, T., González- Roberts, D. A., Dennison, P. E., Gardner, M. E., Hetzel, Y., Prieto, S., & Jiménez, E. (2014). Testing the effects of Ustin, S. L., & Lee, C. T. (2003). Evaluation of the poten- straw mulching and herb seeding on soil erosion after tial of Hyperion for fire danger assessment by compar- fire in a gorse shrubland. Geoderma, 223–225, 79–87. ison to the airborne visible/infrared imaging https://doi.org/10.1016/j.geoderma.2014.01.014 spectrometer. IEEE Transactions on Geoscience and Veraverbeke, S., & Hook, S. J. (2013). Evaluating spectral Remote Sensing, 41(6), 1297–1310. https://doi.org/10. indices and spectral mixture analysis for assessing fire 1109/TGRS.2003.812904 severity, combustion completeness and carbon Rocha, A. V., & Shaver, G. R. (2009). Advantages of a two emissions. International Journal of Wildland Fire, 22(5), band EVI calculated from solar and photosynthetically 707. https://doi.org/10.1071/WF12168 active radiation fluxes. Agricultural and Forest Veraverbeke, S., Lhermitte, S., Verstraeten, W. W., & Meteorology, 149(9), 1560–1563. https://doi.org/10.1016/ Goossens, R. (2011). Evaluation of pre/post-fire differ- j.agrformet.2009.03.016 enced spectral indices for assessing burn severity in Rozario, P., Madurapperuma, B., & Wang, Y. (2018). a Mediterranean environment with Landsat thematic Remote sensing approach to detect burn severity risk mapper. International Journal of Remote Sensing, 32(12), zones in Palo Verde National Park, Costa Rica. Remote 3521–3537. https://doi.org/10.1080/01431161003752430 Sensing, 10(9), 1427. https://doi.org/10.3390/rs10091427 Veraverbeke, S., Verstraeten, W. W., Lhermitte, S., Van De Sarker, S. K., Sonet, S. S., Mohasinul Haque, M., & Kerchove, R., & Goossens, R. (2012). Assessment of post-fire Sharmin, M. (2013). Disentangling the role of soil in changes in land surface temperature and surface albedo, and structuring tropical tree communities at Tarap Hill their relation with fire - burn severity using multitemporal Reserve of Bangladesh. Ecological Research, 28(4), MODIS imagery. International Journal of Wildland Fire, 21 553–565. https://doi.org/10.1007/s11284-013-1045-x (3), 243. https://doi.org/10.1071/WF10075 Sebastián-López, A., Salvador-Civil, R., Gonzalo-Jiménez, J., Vermote, E., Justice, C., Claverie, M., & Franch, B. (2016). & SanMiguel-Ayanz, J. (2008). Integration of Preliminary analysis of the performance of the Landsat 8/ socio-economic and environmental variables for model- OLI land surface reflectance product. Remote Sensing of ling long-term fire danger in Southern Europe. European Environment, 185, 46–56. https://doi.org/10.1016/j.rse. Journal of Forest Research, 127(2), 149–163. https://doi. 2016.04.008 org/10.1007/s10342-007-0191-5 Vivchar, A. (2011). Wildfires in Russia in 2000–2008: Sharma, R. A. (2006). Management plans for Rema-Kalenga Estimates of burnt areas using the satellite MODIS wildlife sanctuary. Dhaka, Bangladesh: Nishorgo Support MCD45 data. Remote Sensing Letters, 2(1), 81–90. Project, Bangladesh Forest Department. https://nishorgo. https://doi.org/10.1080/01431161.2010.499138 org/wp-content/uploads/2017/04/5-5-2-Management_ Vlassova, L., Pérez-Cabello, F., Mimbrero, M., Llovería, R., plans_for_rkws.pdf Smith, C. W., Panda, S. K., Bhatt, U. S., Meyer, F. J., & García-Martín, A. (2014). Analysis of the relationship Badola, A., & Hrobak, J. L. (2021). Assessing Wildfire between land surface temperature and wildfire severity in Burn severity and its relationship with environmental fac- a series of Landsat images. Remote Sensing, 6(7), tors: A case study in interior Alaska Boreal Forest. Remote 6136–6162. https://doi.org/10.3390/rs6076136 Sensing, 13(10), 1966. https://doi.org/10.3390/rs13101966 Wallenius, T. H., Kuuluvainen, T., & Vanha-Majamaa, I. Sobrino, J., Llorens, R., Fernández, C., Fernández-Alonso, J., (2004). Fire history in relation to site type and vegetation & Vega, J. (2019). Relationship between soil burn severity in Vienansalo wilderness in eastern Fennoscandia, in forest fires measured in situ and through spectral Russia. Canadian Journal of Forest Research, 34(7), indices of remote detection. Forests, 10(5), 457. https:// 1400–1409. https://doi.org/10.1139/x04-023 doi.org/10.3390/f10050457 Wendt, C. K., Beringer, J., Tapper, N. J., & Hutley, L. B. Sobuj, N. A., & Rahman, M. (2011). Comparison of plant (2007). Local boundary-layer development over burnt diversity of natural forest and plantations of and unburnt tropical savanna: An observational study. Rema-Kalenga Wildlife Sanctuary of Bangladesh. Boundary-Layer Meteorology, 124(2), 291–304. https:// Journal of Forest Science, 27(3), 127–134. doi.org/10.1007/s10546-006-9148-3 Taylor, C., McCarthy, M. A., & Lindenmayer, D. B. (2014). Wotton, B. M., Martell, D. L., & Logan, K. A. (2003). C. Nonlinear effects of stand age on fire severity. Conservation Climatic Change, 60(3), 275–295. https://doi.org/10.1023/ Letters, 7(4), 355–370. https://doi.org/10.1111/conl.12122 A:1026075919710 Tipu, M., Manzoor Rashid, A., Tahasina, C., & Yang, J., Weisberg, P. J., & Bristow, N. A. (2012). Landsat Mahmood, H. (2021). Semi-destructive approach in remote sensing approaches for monitoring long-term tree developing allometric biomass for chukrasia tabularis in cover dynamics in semi-arid woodlands: Comparison of Sylhet region of bangladesh. Journal of Tropical Forest vegetation indices and spectral mixture analysis. Remote Science, 33(2), 203–212. https://doi.org/10.26525/ Sensing of Environment, 119, 62–71. https://doi.org/10. jtfs2021.33.2.203 1016/j.rse.2011.12.004 Tyukavina, A., Potapov, P., Hansen, M. C., Pickens, A. H., Yin, C., He, B., Yebra, M., Quan, X., Edwards, A. C., Liu, X., Stehman, S. V., Turubanova, S., Parker, D., Zalles, V., & Liao, Z. (2020). Improving burn severity retrieval by Lima, A., Kommareddy, I., Song, X. -P., Wang, L., & integrating tree canopy cover into radiative transfer Harris, N. (2022). Global trends of forest loss due to fire model simulation. Remote Sensing of Environment, 236, from 2001 to 2019. Frontiers in Remote Sensing, 3. https:// doi.org/10.3389/frsen.2022.825190 111454. https://doi.org/10.1016/j.rse.2019.111454 22 MOHAMMED ET AL. Zhao, H. -L., Guo, Y. -R., Zhou, R. -L., & Drake, S. (2010). Zhao, H. -L., Guo, Y. -R., Zhou, R. -L., & Drake, S. (2011). Biological soil crust and surface soil properties in differ- The effects of plantation development on biological soil ent vegetation types of Horqin Sand Land, China. crust and topsoil properties in a desert in northern China. CATENA, 82(2), 70–76. https://doi.org/10.1016/j.catena. Geoderma, 160(3–4), 367–372. https://doi.org/10.1016/j. 2010.05.002 geoderma.2010.10.005
Journal
Geology Ecology and Landscapes – Taylor & Francis
Published: Jan 9, 2023
Keywords: Forest fire; GEE; dNBR; spectral indicies; climate change