Assessment of Pre/Post-wildfire Burned Severity and Soil Erosion Risk for Conservation Planning in Khao Phra District, Thailand Using Remote Sensing and Google Earth Engine

Abstract

Due to changes in climate, non-controlled large-scale wildfires are becoming more frequent and dangerous for human lives and residential areas. A quantitative assessment of soil erosion is needed to outline an evaluation on the extent and magnitude of post-fire soil erosion potential and to develop the effectiveness of the mitigation actions and conservation plans. Thus, the main objective of this study is to assess burned severity and estimate the rate of soil erosion risk by using the coupling of a satellite imagery-based Revised Universal Soil Loss Equation (RUSLE) method and the Google Earth Engine (GEE) platform for the wildfire event on 22-26 February 2020 in Khao Phra district, Thailand. The results of this study indicate that the areas of high and moderate burned severity are ~0.2 (0.36%) and ~11 sq.km. (20.44% of the total of study area), respectively and the soil erosion rates of pre- and post-wildfire are 330 and 459 ton/ha/year, respectively. In addition, our findings reveal significant geographical soil erosion prone locations, which can help in making decisions and developing plans related to the wildfire mitigation actions, humanitarian assistance, and natural conservation.

Introduction

Currently, non-controlled large-scale wildfires are becoming more frequent and dangerous for human lives and residential areas due to climate change. In fact, wildfires impact on communities to disturb critical infrastructure sectors such as transportation, communications, power and gas service, and causing losses of property, crops, resources, animals, and human lives. For long-term harm of wildfires, it can cause for tree and critical root systems, leading to soil erosion, sediment runoff, and contamination concentrations in ecosystems (Mallinis et al., 2013; Davies et al., 2018; Islam et al., 2023; Lee et al., 2019; Han et al., 2017). Naturally, high-intensity rainfall events after a high-severity wildfire can trigger soil erosion. Thus, a plan of post-fire soil erosion control is becoming increasingly necessary to protect environment, communities, and human values at risk.

Hillslope soil erosion after wildfires occurs commonly due to the loss of roots and vegetation holding soil in place and alteration of soil physical and chemical properties. Moreover, severe heating of soils can lead to a reduction in soil infiltration rates and formation of a water-repellent layer in the soil (Deak A., 2023; Garrido-Ruiz et al., 2022). Recently, the modernization of computational tools and remote sensing technology allows the generation of several soil erosion models. The empirical models include e.g. EPM (Gavrilovic, 1962), USLE (Wischmeier and Smith, 1978), RUSLE (Renard et al., 1991) etc., stochastic models e.g. AGNPS (Young et al., 1987), SWAT (Arnold et al., 1998) and deterministic models e.g. ANSWERS (Beasley et al., 1980), WEPP (Nearing et al., 1989), KINEROS (Woolhiser et al., 1990), EUROSEM (Morgan et al., 1998) etc. For applying the soil erosion model of the Revised Universal Soil Loss Equation (RUSLE), several research has been done so far in estimating soil erosion such as the case of Mati fatal wildfire in Eastern Attica, Greece (Efthimiou et al., 2020), the Upper Mad River Basin in California, USA. (Buhr et al., 2016), the case of West Fork wildfire in Colorado, USA. (Yochum and Norman, 2015), the Basilicata region, Italy (Lanorte, et al., 2019), the Puglia region, Italy (Nole et al., 2020), and the Upper Ping River Watershed in Thailand (Rapeepong, 2014). The Google Earth Engine (GEE) is one of cloud-based platforms for the Application Program Interface (API). The GEE’s databases include many different data of satellite images, geospatial datasets, meteorological data, land cover/land use data, topographic data, and environmental data that have the advantages of a wide observation range and low cost to achieve long-term series monitoring of soil erosion. Digital elevation model, land use, and weather parameters can be extracted conveniently and economically from satellite imagery for the soil erosion model (GEE, 2023). Therefore, the main objective of this study is to assess burned severity and estimate the rate of soil erosion risk by using the coupling of a satellite imagery-based Revised Universal Soil Loss Equation (RUSLE) method and the Google Earth Engine (GEE) platform for pre- and post-wildfire on 22-26 February 2020 in Khao Phra district, Thailand. In fact, recent research has not established a systematic technical method for monitoring and assessing the effects of wildfires in Khao Phra district. It is necessary to conduct in-depth research on the conservation plan with a focus on exploring effective technical methods to help in making decisions and developing plans related to the wildfire mitigation actions, humanitarian assistance, and natural conservation. Moreover, this research can support for achieving the United Nations Sustainable Development Goals (SDGs), target 3 and goal 11, which is to provide communities with safer, healthier, and more prosperous lifestyles by allowing forests to reduce climate change migration (Turner-Skoff and Cavender, 2019).

Methodology

Study Area

The study area of Khao Phra district which is located at 101°19´and 101°22´east longitudes and between 14°25´and 14°31´north latitudes with 728 sq.km. area in Nakhonnayok province of Thailand (Nakhonnayok, 2023). The location of study area is showed in Figure 1. Characteristics of the land of Khao Phra district are mountainous which is covered by the full of Khao Laem and Khao Phra mountains. The maximum temperature of Khao Phra district is ~34 °C in April and the minimum temperature is ~24 °C in December. The district’s annual average precipitation is reported at 193 mm. The number of populations in Khao Phra district is 8,565 people (Nakhonnayok, 2023).

Figure 1. The map of study area location

Data and Conceptual Framework

In this study, pre- and post-wildfire burned severity and soil erosion risk are estimated and evaluated for planning a conservation in the study area. For assessing and generating the burn severity map, sentinel 1 SAR VV, 2A and 2B satellite datasets of pre- and post-wildfire are downloaded through the Google Earth Engine (GEE) and consist of 13 spectral bands of varying spatial resolutions from 10 m to 60 m. The spectral bands are pre-processed to get the geometric corrected TOA reflectance bands. In this study, Bands 3, 8, and B12 are used to generate a map of the burned severity area by calculating the Normalized Burn Ratio (NBR). Figure 2 shows the methodology of burned severity assessment.

Figure 2. The methodology of burned severity assessment

For assessing soil erosion risk, the Revised Universal Soil Loss Equation (RUSLE) model is used in this study. This RUSLE model require five factors for the model including rainfall erosivity (R), soil erodibility (K), slope length and steepness (LS), crop management (C), and support practices (P) (Renard et al., 1991). Table 1 presents the datasets used in the study and Figure 3 shows the conceptual framework of soil erosion risk assessment using Google Earth Engine (GEE) platform.

Table 1. Datasets of RUSLE model in this study

Factor	Data	Dataset provider	Scale	Unit
Rainfall erosivity (R)	Rainfall	U.S. Geological Survey (USGS) of Earth Resources Observation and Science (EROS) Center	global, 0.05°	MJ mm ha-1 hr-1 year-1
Soil erodibility (K)	Soil texture	U.S. Department of Agriculture (USDA)	global, 250 m	tons hr ha-1 MJ-1 mm-1
Slope length and steepness (LS)	Digital Elevation Model	U.S. Geological Survey (USGS)	global, 30 m	dimensionless
Crop management (C)	Normalize Difference Vegetation Index	European Space Agency	global, 10 m	dimensionless
Support practices (P)	Land cover	U.S. Geological Survey (USGS)	global, 30 m	dimensionless

Figure 3. The conceptual framework of soil erosion risk assessment using Google Earth Engine (GEE) platform

Method of Data Analysis

The Normalized Burn Ratio (NBR) and The Differenced Normalized Burn Ratio (dNBR)

The Normalized Burn Ratio (NBR) and the differenced Normalized Burn Ratio (dNBR) are used to assess and generate a map of burn severity in the study area. The algorithm of NBR consists of the near-infrared (NIR) and short-wave infrared (SWIR) bands from satellite-imaged sensor. This algorithm is as follows (Suresh et al., 2018):

  

$$NBR=\frac{(NIR-SWIR)}{(NIR+SWIR)}$$(1)

where NBR is Normalized Burn Ratio, NIR is near-infrared bands, and SWIR is short-wave infrared bands. Commonly, NBR values are ranging from -1 to +1. In NIR region, healthy vegetation has very high reflectance and the burned areas show low reflectance. In contrast, burned areas show higher reflectance and healthy vegetation presents lower in SWIR region. Thus, the higher NBR value means healthy vegetation and lower value of burned areas.

The dNBR is to calculate for the burned severity areas (Kokaly et al., 2007). The dNBR can be estimated from the NBR of pre- and post-wildfire events as shown in equation (2).

  

$$dNBR={NBR}_{pre-fire}-{NBR}_{post-fire}$$(2)

where dNBR is the differenced Normalized Burn Ratio, ${NBR}_{pre-fire}$ is the Normalized Burn Ratio of pre-wildfire event, and ${NBR}_{post-fire}$ is the Normalized Burn Ratio of post-wildfire event. According to the United States Geological Survey (USGS) (UN, 2020), burn severity map is reclassified in 7 interpretation levels based on the dNBR value as shown in Table 2.

Table 2. Burn severity level based on USGS

No.	Severity level	dNBR values
1	Enhanced Regrowth, High	-500 to -251
2	Enhanced Regrowth, Low	-250 to -101
3 4 5 6 7	Unburned Low Severity Moderate-Low Severity Moderate-High Severity High Severity	-100 to +99 +100 to +269 +270 to +439 +440 to +659 +660 to +1300

The Revised Universal Soil Loss Equation (RUSLE)

In this study, the Revised Universal Soil Loss Equation (RUSLE) model is selected for estimating soil erosion risk in the Google Earth Engine (GEE) cloud-based platform for assessing the pre- and post-wildfires event on 22-26 February 2020 in Khao Phra district of Thailand. The RUSLE model integrated with GEE platform is used to predict average annual soil erosion. The RUSLE model is empirically expressed as (Renard et al., 1991):

A= R× K× LS× C× P (3)

where A is the average annual soil erosion per unit area (tons ha^-1 year^-1), R is the rainfall erosivity factor (MJ mm ha^-1 hr^-1 year^-1), K is the soil erodibility factor (tons hr ha^-1 MJ^-1 mm^-1), LS is the slope length-steepness factor (dimensionless), C is the cover and management factor (dimensionless), and P is the erosion support practice or land management factor (dimensionless).

Rainfall Erosivity Factor (R-factor)

The R factor quantifies the effect of rainfall impact and reflects the amount and rate of runoff by considering from the intensity and duration of rainfall (Stone and Hilborn, 2000). The Land Development Department of Thailand (LDD) recommends the following equation for calculating the rainfall erosivity factor (R) as follows (LDD, 2020):

$R=0.4669X-12.1415$ (4)

where R is the rainfall erosivity factor (MJ mm ha^-1 hr^-1 year^-1) and X is the average annual rainfall (mm). In this study, the average annual rainfall is used from the Climate Hazards Group InfraRed Precipitation with Station data (CHIRPS). The CHIRPS data is a quasi-global rainfall dataset, which incorporates 0.05° resolution satellite imagery with in-situ station data to create gridded rainfall time series for trend analysis and seasonal drought monitoring by scientists at the U.S. Geological Survey (USGS) of Earth Resources Observation and Science (EROS) Center since 1999 (Funk et al., 2015). The data of the CHIRPS for this study is obtained from the Google Earth Engine (GEE) with the dataset in a global format and clipped with the boundary of study area.

Soil Erosivity Factor (K-factor)

The K factor quantifies the cohesive character of a soil type and its resistance to detachment and transport due to raindrop impact and runoff (Stone and Hiborn, 2000). Thus, the K factor depends on the physical and chemical properties of the soil, which contribute to its erodibility potential. In this study, soil texture data is obtained from the U.S. Department of Agriculture (USDA). The USDA soil data is derived from predicted soil texture fractions at 250 m using the soil texture package in R (Hengl, 2018). The data of USDA soil texture is available for downloading from the GEE with the dataset in a global format. For K values, the study assigned to each soil texture class based on the Land Development Department of Thailand (LDD) (LDD, 2020). Table 3 shows K values in each soil texture class for the study.

Table 3. K values in each soil texture class for this study (LDD, 2020)

No.	Soil texture	Abbreviation	K value
			Khao Phra
1	Sand	Sa	0.04
2	Loamy sand	LoSa	0.04
3	Sandy loam	SaLo	0.29
4	Loam	Lo	0.29
5	Silt loam	SiLo	0.37
6	Silt	Si	0.57
7	Sandy clay loam	SaClLo	0.24
8	Clay loam	ClLo	0.25
9	Silty clay loam	SiClLo	0.46
10	Sand clay	SaCl	0.15
11	Silty clay	SiCl	0.23
12	Clay	Cl	0.13

Slope Length-steepness Factor (LS-factor)

The LS factor consists of L is the slope length of the land surface, and S is the slope steepness, which affects soil loss and sediment displacement. The slope-length factor (L) is the ratio of soil loss from a field slope length to that from a slope 22.13 m long, and it measures the distance from the origin of overland flow along the flow path to the location of deposition. The slope-steepness factor (S) is the ratio of soil erosion from the field slope gradient to that from a 9% slope (Wischemier and Smith, 1978). In this study, the digital elevation model (DEM) of Shuttle Radar Topographic Mission (SRTM) with a 30 m spatial resolution is used for the calculation of the LS factor (Farr et al., 2007). The data of DEM is available for downloading from the GEE with the dataset in a global format and clipped with the boundary of study area. The study uses the following equation of Wischmeier and Smith (1978) for the calculation of the LS factor:

$LS={\left(\frac{l}{72.6}\right)}^m(65.41{sin}^2\theta +4.56sin\theta +0.065)$ (5)

where l is the cumulative slope length in meter, θ is the downhill slope angle and m is the slope contingent variable.

Cover and Management Factor (C-factor)

The C factor represents the ratio of soil erosion from the effects of cropping and management practices. In this study, the Normalized Difference Vegetation Index (NDVI) is used for the calculation of the C factor. The NDVI is an effective remote sensing indicator of green vegetation distribution that is measured from the difference of the spectral reflectance values between the near-infrared (NIR) and red (RED) bands of the electromagnetic spectrum (Rouse et al., 1974). Basically, the NDVI value varies between -1 to +1 and can be calculated as follows:

$NDVI=\frac{NIR-RED}{NIR+RED}$ (6)

where NIR is the spectral reflectance value of near-infrared light, which vegetation strongly reflects and RED is the spectral reflectance value of red light, which vegetation absorbs. In this study, the NDVI map is generated by performing image classification on the Sentinel-2 level-1C data. The NDVI-values are scaled to estimate C-values in the range of 0-1 using the following formula (Van der Knijff el al., 1999):

$C=exp[-\alpha •\frac{NDVI}{(\beta -NDVI)}]$ (7)

where $\alpha$ , $\beta$ is the parameters that determine the shape of the NDVI-C curve. Theoretically, an α-value of 2 and a β-value of 1 are determined by Van der Knijff et al. (1999).

Support Practice Factor (P-factor)

The P factor quantifies the impact of specific agricultural support practices on the soil loss with upslope and downslope tillage around the watershed. In this study, the P factor is calculated by assigning appropriate values to the Land Use and Land Cover (LULC) classes based on the Land Development Department of Thailand (LDD) (LDD, 2020). The LULC map is obtained from evaluating Moderate-resolution Imaging Spectroradiometer (MODIS) data. The data of LULC is available for downloading from the GEE with the dataset in a global format and clipped with the boundary of study area. Table 4 shows P values in each LULC class for the study.

Table 4. P values in each land use and land cover class for this study (LDD, 2020)

No.	LULC class	P value
1	Dry evergreen forest	0.1
2	Deciduous dipterocarp forest	1.0
3	Mixed deciduous forest	1.0
4	Forest plantation	1.0
5	Grass land	1.0
6	Paddy field	0.1
7	Field crop	1.0
8	Urban area	0
9	Water body	0

Class of Soil Erosion Risk Assessment

The calculated soil erosion values from equation (1) are categorized into 5 classes of soil erosion risk assessment as presented in Table 5, which are proposed by the Land Development Department of Thailand (LDD) (LDD, 2020).

Table 5. Class of soil erosion risk assessment (LDD, 2020)

Class	Soil erosion risk assessment	Rate of soil erosion
		ton/ha/year
1	Very mild	0.00 – 12.5
2	Mild	12.50 – 31.25
3	Moderate	31.25 – 93.75
4	Severe	93.75 -125
5	Very severe	>125

Accuracy Assessment

For the accuracy assessment of calculated soil erosion values, the R-squared (R²) is used to assess the goodness of fit of the model with the R² values between 0 and 1. The calculation of R² is described as the following equation:

$R^2=1-\frac{{\sum }_{}^{}{(y_i-{\hat{y}}_i)}^2}{{\sum }_{}^{}{(y_i-\overline{y})}^2}$ (8)

where $y_i$ is the observed value, ${\hat{y}}_i$ is the predicted value, and $\overline{y}$ is the mean of the y values. In this study, the relationship between soil erosion rates from the RUSLE model and sediment loading rates from the Royal Irrigation Department (RID) (RID, 2019) is evaluated and considered on the values of R².

Results and Discussion

Burned Severity Assessment

After the wildfire event on 22-26 February 2020 in Khao Phra district of Thailand, sentinel 2A and 2B satellite datasets are used to generate the burn severity map. The results show that the area of high severity is ~0.2 sq.km. (0.36% of the total of study area), moderate severity is ~11 sq.km. (20.44% of the total of study area), low severity is ~16 sq.km. (29.42% of the total of study area), and unburned area is ~25 sq.km. (45.25% of the total of study area) as shown in Figure 4. It should be noted that the area of high severity occurred mostly in the southern part of study area where is near many residential areas, e.g. the Armed Forces Academies Preparatory School. Thus, this high-severity area should be planned significantly for the wildfire protection and mitigation actions such as creating wildfire barriers etc. In comparison with Juntakut et al. (2021) and Juntakut et al. (2023), the study demonstrated the burned area resulting from wildfire events in the region. The findings clearly showed that the most severe damage to facilities occurred in the southern part of the study area, aligning with the results of this research.

Figure 4. Burned severity map on the study area using Google Earth Engine (GEE)

Pre- and Post-wildfire Soil Erosion Risk Assessment

The rates of soil erosion from the pre- and post-wildfire event are estimated and assessed at risk using RUSLE model in GEE as shown in Figure 5 and Figure 6. The results indicate that the soil erosion rates of pre- and post-fire in Khao Phra district are 330 and 459 ton/ha/year, respectively. For the assessment of soil erosion risk, the area in the level of very severe (>125 ton/ha/year) is ~47% of the total of study area for pre-wildfire event. In contrast, the area in the level of very severe is higher as ~57% of the total of study area for post-wildfire event. Significantly, it is ~10% more areas of soil erosion at risk after the wildfire event. As a result, the significant information of geographical soil erosion prone locations can help as an information in making decisions and developing plans related to the wildfire mitigation actions, humanitarian assistance, and natural conservation in Khao Phra district.

Figure 5. Assessing pre-wildfire soil erosion risk on the study area using Google Earth Engine (GEE) with the result of ~47% of the total of study area under the level of very severe in soil erosion at risk

Figure 6. Assessing post-wildfire soil erosion risk on the study area using Google Earth Engine (GEE) with the result of ~57% of the total of study area under the level of very severe in soil erosion at risk

Assessment of Soil Erosion Rates

Figure 7 show the relationship between soil erosion rates in Khao Phra district from RUSLE model and sediment loading rates from the Royal Irrigation Department (RID). Data of sediment loading rates is collected and reported by the RID (RID, 2019). In the study, these sediment loading rates are used for estimating R² of the relationship from four stations around the area of Khao Phra district. The results demonstrate that the relationship of R² is 0.82. This R² value can explain about 82% of total variation of the sediment loading rates by using this power relationship. In comparison with Juntakut et al. (2021) and Juntakut et al. (2023), this study highlighted the soil erosion rates resulting from wildfire events in the region, which are consistent with the findings of this research.

Figure 7. Relationship between soil erosion rates in the study area from RUSLE model in Y axis and sediment loading rates from the Royal Irrigation Department in X axis.

Conclusion

In this study, pre- and post-wildfires burned severity and soil erosion risk are estimated and assessed for planning a conservation. For generating the burn severity map from the wildfire event on 22-26 February 2020 in Khao Phra district of Thailand, sentinel 1 SAR VV, 2A and 2B satellite datasets of pre- and post-wildfire are used through the Google Earth Engine (GEE) to analyse the burned severity area by calculating the Normalized Burn Ratio (NBR) and the differenced Normalized Burn Ratio (dNBR). The results show that the area of high severity is ~0.2 sq.km. (0.36%), moderate severity is ~11 sq.km. (20.44%), low severity is ~16 sq.km. (29.42%), and unburned area is ~25 sq.km. (45.25% of the total of study area). It should be noted that the area of high severity occurred mostly in the southern part of study area where is near many residential areas in the study area. Therefore, it is necessary to plan significantly for the wildfire protection and mitigation actions such as creating wildfire barriers etc. to slow or stop the spread of wildfire that it helps protect the communities from catching fire.

The rates of soil erosion from the pre- and post-wildfire event are estimated and assessed at risk using RUSLE model in GEE. The results indicate that the soil erosion rates of pre- and post-fire in Khao Phra district are 330 and 459 ton/ha/year, respectively. For the assessment of soil erosion risk, the area in the level of very severe (>125 ton/ha/year) is ~47% of the total of study area for pre-wildfire event. In contrast, the area in the level of very severe is higher as ~57% of the total of study area for post-wildfire event. In the accuracy assessment of the results, the relationship between soil erosion rates in Khao Phra district from RUSLE model and sediment loading rates from the Royal Irrigation Department (RID) is estimated with R². The result of R² is 0.82 (82%). As a result, the significant information of geographical soil erosion prone locations can help in making decisions and developing plans related to the wildfire mitigation actions, humanitarian assistance, and natural conservation in Khao Phra district. In addition, this study can also support for achieving the United Nations Sustainable Development Goals (SDGs), target 3 and goal 11, which is to provide communities with safer, healthier, and more prosperous lifestyles by allowing forests to reduce climate change migration.

Author Contributions

Conceptualization, J.P., J.Y. and J.C.; methodology, J.P. and J.Y.; software, J.P.; investigation, J.P., J.Y. and J.C.; resources, J.P.; data curation, J.P.; writing—original draft preparation, J.P., J.Y. and J.C.; writing—review and editing, J.P., J.Y. and J.C. All authors have read and agreed to the published version of the manuscript.

Ethics Declaration

The authors declare that they have no conflicts of interest regarding the publication of the paper.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

References

•  Arnold , J.G.,  Srinivasan,  R., et al. (1998). “Large area hydrologic modeling and assessment part I: model development”, Journal of the American Water Resources Association, 34:73–89. doi: https://doi.org/10.1111/j.1752-1688.1998.tb05961.x.
•  Beasley , D.B.,  Huggins  L.F., et al. (1980). “ANSWERS: A model for watershed planning”. Transactions of the ASAE, 23:938–944.
•  Buhr,  D.,  Camplair,  M., et al. (2016). “Modeling Post-Fire Sediment Yields and Reservoir Capacity in the Upper Mad River Basin, California”, Retrieved from https://agsci.oregonstate.edu/sites/agscid7/files/eisi/buhr_camplair_hamilton_eisi_2016_paper_1.pdf on Mar. 09, 2024.
•  Davies , I.P.,  Haugo , R.D., et al. (2018). “The unequal vulnerability of communities of color to wildfire”, PLoS ONE 13(11): e0205825. doi: https://doi.org/10.1371/journal.pone.0205825.
•  Deak,  F. (2023). “Post-Fire Erosion Control”, University of California Cooperative Extension Mariposa, Madera, and Fresno Counties.
•  Efthimiou,  N.,  Psomiadis,  E., et al. (2020). “Fire severity and soil erosion susceptibility mapping using multi-temporal Earth Observation data: The case of Mati fatal wildfire in Eastern Attica, Greece”, Catena, 187, 104320. doi: https://doi.org/10.1016/j.catena.2019.104320.
•  Farr , T.G.,  Rosen , P.A., et al. (2007). “The shuttle radar topography mission”, Reviews of Geophysics, v. 45, no. 2, RG2004. doi: https://doi.org/10.1029/2005RG000183.
•  Funk,  C.,  Peterson,  P., et al. (2015). “The climate hazards infrared precipitation with stations-a new environmental record for monitoring extremes”, Scientific Data 2, 150066. doi: https://doi.org/10.1038/sdata.2015.66.
•  Garrido-Ruiz,  C.,  Sandoval,  M., et al. (2022). “Fire impacts on soil and post fire emergency stabilization treatments in Mediterranean-climate regions”, Chilean Journal of Agricultural Research, 82(2), 335-347.
• Gavrilovic, S. (1962). “A method for estimating the average annual quantity of sediments according to the potency of erosion”. Bullet. Faculty Forestry, 26:151–168.
•  GEE . (2023). “Google Earth Engine”, Retrieved from https://earthengine.google.com/ on Mar. 31, 2023
•  Han,  H.,  Ma,  Q., et al. (2017). “Urban Growth Boundaries of the Hangzhou Metropolitan Area Based on Ecosystem Service”, International Review for Spatial Planning and Sustainable Development, 5(2), 4-16. doi: https://doi.org/10.14246/irspsd.5.2_4.
•  Hengl,  T. (2018). “Soil texture classes”, (USDA system) for 6 soil depths (0, 10, 30, 60, 100 and 200 cm) at 250 m (Version v02).
• Islam, M.S., Teasmin, T., et al. (2023). “Vegetation Cover Change Analysis during 1989-2020 of Coastal Barguna District, Bangladesh Using Remote Sensing and GIS Technology”, International Review for Spatial Planning and Sustainable Development, 11(2), 259-277. doi: https://doi.org/10.14246/irspsd.11.2_259.
•  Juntakut,  P.,  Jantakat,  Y., et al. (2021). “Guideline of The Application of Google Earth Engine for Monitoring and Damage Assessment of Natural Disaster”, the 26^th National Convention on Civil Engineering.
•  Juntakut,  P.  Jantakat,  Y., et al. (2023). “The BOOG-FAG Web App – A Decision-Support Tool for Soil Erosion Risk Assessment and Mitigation after Wildfires in Case Study of Khao Laem, Thailand”, The 18^th APRU Multi-Hazards Symposium 2023 in Singapore.
•  Kokaly , R.F.,  Rockwell , B.W., et al. (2007). “Characterization of post-fire surface cover, soils, and burn severity at the Cerro Grande Fire, New Mexico, using hyperspectral and multispectral remote sensing”, Remote Sensing of Environment, 106(3), 305-325. doi: https://doi.org/10.1016/j.rse.2006.08.006.
•  Lanorte,  A.,  Cillis,  G., et al. (2019). “Integrated approach of RUSLE, GIS and ESA Sentinel-2 satellite data for post-fire soil erosion assessment in Basilicata region (Southern Italy)”, Geomatics, Natural Hazards and Risk, 10(1), 1563-1595.
•  LDD . (2020). “Land Development Department: Status of Soil Erosion in Thailand”, Retrieved from https://www.ldd.go.th/Web_Soil/PDF/soilerosion.pdf on Mar. 30, 2023.
•  Lee , J.H.,  Kim , S.H., et al. (2019). “Boundary Setting for Ecosystem Services by Factor Analysis”, International Review for Spatial Planning and Sustainable Development, 7(2), 21-35. doi: https://doi.org/10.14246/irspsd.7.2_21.
•  Mallinis,  G.,  Maris,  F., et al. (2013). “Assessment of Postfire Soil Erosion Risk in Fire-Affected Watersheds Using Remote Sensing and GIS”, GIScience and Remote Sensing, 46:4, 388-410. doi: https://doi.org/10.2747/1548-1603.46.4.388.
•  Morgan , R.P.C.,  Quinton , J.N., et al. (1998). “The European Soil Erosion Model (EUROSEM): A dynamic approach for predicting sediment transport from fields and small catchments”. Earth Surface Processes and Landforms, 23:527–544.
•  Nakhon  Nayok. (2023). “Tourism Authority of Thailand (TAT)”, Retrieved from https://web.archive.org/web/20150810120952 on Jun. 25, 2023.
•  Nearing , M.A.,  Foster , G.R., et al. (1989). “A process-based soil erosion model for USDA: water erosion prediction project technology”, Transactions of the ASAE, 32:1587–1593.
•  Nole,  G.,  Santarsiero,  V., et al. (2020). “Model of Post Fire Erosion Assessment Using RUSLE Method, GIS Tools and ESA Sentinel DATA”, Springer, ICCSA 2020, LNCS 12253, 505–516.
•  Rapeepong,  L. (2014). “Assessment of soil loss in the Upper Ping River Basin”, Bangkok: Department of Water Resources Engineering, Kasetsart university.
•  Renard , K.G.,  Foster , G.R., et al. (1991). “RUSLE: Revised Universal Soil Loss Equation”. Journal of Soil and Water Conservation, 46(1), 30–33.
• RID. (2019). “Royal Irrigation Department: The Relation between Suspended Sediment and Drainage Area in 25 River Basins”, Retrieved from http://hydrology.rid.go.th/sediment-wq/filepdf/suspendsediment/suspended%20sediment%2025%20basin_62.pdf on Mar. 20, 2023.
•  Rouse , J.W.,  Haas , R.W., et al. (1974). “Monitoring the vernal advancement and retrogradation (greenware effect) of natural vegetation”, Greenbelt, MD, USA, NASA/GSFCT, Type 3, Final Report.
•  Stone , R.P., and  Hilborn,  D. (2000). “Universal Soil Loss Equations Fact Sheet (USLE)”, Ministry of Agriculture, Food and Rural Affairs Version 2 (RUSLE2).
•  Suresh , B.K.V.,  Arijit,  R., et al. (2018). “Mapping of Forest Fire Burned Severity using the Sentinel Datasets”, The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XLII-5.
• Turner‐Skoff, J.B., and Cavender, N. (2019). “The benefits of trees for liveable and sustainable communities”, PLANTS, PEOPLE, PLANET, 1(4), 323–335. doi: https://doi.org/10.1002/ppp3.39.
• UN. (2020). “UN-SPIDER”, Retrieved from https://www.un-spider.org/advisory-support/recommended-practices/recommended-practice-burn-severity/in-detail/normalized-burn-ratio on Jun. 20, 2023.
•  Van der Knijff , J.M.,  Jones , R.J.A., et al. (2000). “Soil erosion risk assessment in Italy”, European Soil Bureau, Joint Research Center of the European Commission. In press.
•  Wischmeier , W.H. and  Smith , D.D. (1978). “Predicting Rainfall Erosion Losses. A Guide to Conservation Planning”. USDA Agric. HB No. 537. Washington DC.
•  Woolhiser , D.A.,  Smith , R.E., et al. (1990). “A kinematic runoff and erosion model”, Documentation and user manual USDA-ARS-77. Washington DC.
•  Young , R.A.,  Onstad , C.A., et al. (1987). “USDA Conservation”, Research Report No.35. Washington DC. 1987. AGNPS, Agricultural non-Point-Source Pollution Model. A watershed analysis tool.