Development of a Smoke Dispersion Forecast System for Korean Forest Fires

Lee, Boknam; Cho, Seungwan; Lee, Seung-Kii; Woo, Choongshik; Park, Joowon

doi:10.3390/f10030219

Cited by 10 publications

(5 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Different methods of estimating emissions are discussed in the literature, namely inversion methods coupled with Gaussian dispersion models (Krings et al, 2011;Nassar et al, 2017;Lee et al, 2019), different chemical transport models (CTMs) (Brasseur and Jacob, 2017), cross-sectional flux method (CFM) (White et al, 1976;Beirle et al, 2011;Cambaliza et al, 2014Cambaliza et al, , 2015Kuhlmann et al, 2020), and integrated mass enhancement (IME) method (Frankenberg et al, 2016). An inversion coupled with a Gaussian plume model is used for flux inversions of an isolated single plume assuming steady and uniform wind conditions.…”

Section: Introductionmentioning

confidence: 99%

Plume detection and emission estimate for biomass burning plumes from TROPOMI carbon monoxide observations using APE v1.1

Goudar,

Anema,

Kumar

et al. 2023

Geosci. Model Dev.

View full text Add to dashboard Cite

Abstract. This paper presents the automated plume detection and emission estimation algorithm (APE), developed to detect CO plumes from isolated biomass burning events and to quantify the corresponding CO emission rate. APE uses the CO product of the Tropospheric Monitoring Instrument (TROPOMI) on board the Copernicus Sentinel-5 Precursor (S5P) satellite, launched in 2017, and collocated active fire data from the Visible Infrared Imaging Radiometer Suite (VIIRS), the latter flying 3 min ahead of S5P. After identifying appropriate fire events using VIIRS data, an automated plume detection algorithm based on traditional image processing algorithms selects plumes for further data interpretation. The approach is based on thresholds optimized for data over the United States in September 2020. Subsequently, the CO emission rate is estimated using the cross-sectional flux method, which requires horizontal wind fields at the plume height. Three different plume heights were considered, and the ECMWF Reanalysis v5 (ERA5) data were used to compute emissions. A varying plume height in the downwind direction based on three-dimensional Lagrangian simulation was considered appropriate. APE is verified for observations over Australia and Siberia. For all fire sources identified by VIIRS, only 16 % of the data corresponded to clear-sky TROPOMI CO data with plume signature. Furthermore, the quality filters of APE resulted in emission estimations for 26 % of the TROPOMI CO data with plume signatures. Visual filtering of the APE's output showed a true-positive confidence level of 97.7 %. Finally, we provide an estimate of the emission uncertainties. The greatest contribution of error comes from the uncertainty in Global Fire Assimilation System (GFAS) injection height that leads to emission errors <100 %, followed by systematic errors in the ERA5 wind data. The assumption of constant emission during plume formation and spatial under-sampling of CO column concentration by TROPOMI yields an error of <20 %. The randomized errors from the ensemble ERA5 wind data are found to be less than 20 % for 97 % of the cases.

show abstract

Section: Introductionmentioning

confidence: 99%

Plume detection and emission estimate for biomass burning plumes from TROPOMI carbon monoxide observations using APE v1.1

Goudar,

Anema,

Kumar

et al. 2023

Geosci. Model Dev.

View full text Add to dashboard Cite

show abstract

“…Risk prediction studies for disaster preparation and prevention have been conducted in various fields, such as forest fire prediction (5)(6)(7)(8) and crime prediction. (9,10) In urban fire prediction, Wang et al (11) performed a grid-scale spatiotemporal prediction using the combining gate recurrent unit and conditional random field (GRU-CRF) model to predict fire risk in the Zhengzhou area.…”

Section: Introductionmentioning

confidence: 99%

Grid-based Urban Fire Prediction Using Extreme Gradient Boosting (XGBoost)

Oh¹,

Jeong²

2022

Sensors and Materials

View full text Add to dashboard Cite

Fires in urban areas lead to enormous financial and human losses because cities have high densities of people and buildings. Although a recent advanced IoT technology improves early fire detection, it is crucial to predict fire risk to manage and prevent urban fires. We propose a method of predicting urban fires using extreme gradient boosting (XGBoost), which is based on grid-based data, to consider the characteristics of urban fires occurring in local areas. Before model training, we conducted a correlation analysis and variance inflation factor (VIF) analysis to remove variables with a strong correlation between independent variables. Furthermore, oversampling and feature selection techniques were applied to improve the model's performance. Experimental results revealed that the overall accuracy of XGBoost was 81.25%, the F1-score was 86.43%, and the area under the curve (AUC) was 84.59%. XGBoost performed better than baseline models, such as the support vector machine (SVM) and logistic regression. The results of this study show that it can be used for local area management and the prevention of urban fires.

show abstract

“…The CO plumes in the TROPOMI data can be used to estimate CO emission by wild fires and different methods are discussed in the literature, namely, the inversion methods coupled with Gaussian dispersion models (Krings et al, 2011;Nassar et al, 2017;Lee et al, 2019) or different Chemical Transport Models (CTM) (Brasseur and Jacob, 2017), Cross-sectional Flux Methods (CFM) (White et al, 1976;Beirle et al, 2011;Cambaliza et al, 2014Cambaliza et al, , 2015Kuhlmann et al, 2020) and integrated mass enhancement (IME) method (Frankenberg et al, 2016). The inversion coupled with a Gaussian plume model is simple where an analytically computed Gaussian plume is fitted to TROPOMI CO column observations.…”

Section: Introductionmentioning

confidence: 99%

Plume detection and estimate emissions for biomass burning plumes from TROPOMI Carbon monoxide observations using APE v1.0

Goudar

Anema

Kumar

et al. 2022

Preprint

View full text Add to dashboard Cite

Abstract. The TROPOspheric Monitoring Instrument (TROPOMI) onboard the Sentinel-5 Precursor (S-5P) satellite, launched in 2017, measures the total column concentration of the trace gas Carbon Monoxide (CO) daily on a global scale and at a high spatial resolution of 7 x 7 km2, improved to 5.5 x 7 km2 in August 2019. The TROPOMI observations show plumes of CO due to localized CO emissions from industrial sources and biomass burning. In this paper, we quantify these CO emissions for biomass burning by an automated algorithm, APE, to detect plumes and quantify the CO emission rate using cross-sectional flux method. Furthermore, the influence of a constant and a varying plume height in downwind direction on emissions is investigated and algorithm uncertainties are quantified. The VIIRS active fire data in conjunction with the TROPOMI CO datasets is used to identify fires and the fire locations. Then, an automated plume detection algorithm using traditional image processing algorithms is developed and utilized to identify plumes. For these plumes, the emission rate is estimated by the cross-section flux method at three different plume heights. The first two are constant plume heights at a 100 m and an IS4FIRES injection height from Global Fire Assimilation System. And the last one is a varying plume height in downwind direction. A 3D Lagrangian model is used to simulate tracer particles where the source locations for the simulation are based on the VIIRS fire counts and IS4FIRES injection height. 3D velocities at 137 model levels (ERA5) are utilized to simulate tracer particles. We demonstrate the quality and validity of our automated approach by investigating biomass burning events and their emissions for Australia on Oct 2019 and the US on Sept 2020. A total of 110 and 31 individual fire plumes in Australia and the US, respectively were detected and their emissions estimated. The emissions were severely under-predicted and negative for 11 cases when based on constant plume height of 100 m compared to emissions based on varying plume height. Furthermore, the effect of the changing plume height in downwind direction on the emission estimate compared to emissions from constant IS4FIRES plume height was minor as 124 cases are found to have emission variation less than 10 %. However, we were able to identify several cases where the flux estimates become more reliable with varying plume height. Thus, the varying plume height in downwind direction is considered for the automated algorithm. The cross-section flux method is found to have an uncertainty of 38 % in one of the idealized cases. However, overall uncertainty of the algorithm is difficult to quantify as conditions for each fire are unique.

show abstract

Development of a Smoke Dispersion Forecast System for Korean Forest Fires

Cited by 10 publications

References 27 publications

Plume detection and emission estimate for biomass burning plumes from TROPOMI carbon monoxide observations using APE v1.1

Plume detection and emission estimate for biomass burning plumes from TROPOMI carbon monoxide observations using APE v1.1

Grid-based Urban Fire Prediction Using Extreme Gradient Boosting (XGBoost)

Plume detection and estimate emissions for biomass burning plumes from TROPOMI Carbon monoxide observations using APE v1.0

Contact Info

Product

Resources

About