Sainbayar Dalantai
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Sainbayar Dalantai has a total of 41 co-authors. Here are the top co-authors by number of shared publications:
Amarsaikhan Damdinsuren
Division of Aerospace Remote Sensing, Institute of Geography and Geoecology, Mongolian Academy of Sciences, Ulaanbaatar, Mongolia
4 shared publications- Amarsaikhan Damdinsuren
Institute of Geography and Geoecology, Mongolian Academy of Sciences, Ulaanbaatar 15170, Mongolia
3 shared publications - Byambadolgor Batdorj
Institute of Geography and Geoecology, Mongolian Academy of Sciences, Ulaanbaatar 15170, Mongolia
3 shared publications Bayartungalag Batsaikhan
Department of Earth and Environmental Sciences Environmental Geochemistry Laboratory Korea University Seoul Republic of Korea
2 shared publications- Boldbaatar Natsagdorj
Department of Environment and Forest Engineering, School of Engineering and Technology, National Univesity of Mongolia, Ulaanbaatar 15170, Mongolia
2 shared publications - Erdenesukh Sumiya
National University of Mongolia (NUM), Ulaanbaatar, Mongolia
2 shared publications
Latest Publications
- Open Access
- 6 April 2026
- Published by Mongolian Journals Online in Mongolian Geoscientist
- Vol. 31 (62) , 99-114
Green vegetation biomass estimation in Mongolia using RS data and machine-learning methodsing methods
Green vegetation biomass is a key ecological indicator reflecting ecosystem productivity, rangeland condition, and responses to climate variability, particularly in arid and semi-arid environments. In Mongolia, where vast grassland and desert-steppe ecosystems support both biodiversity and pastoral livelihoods, spatially explicit biomass information is essential for ecological monitoring and conservation-oriented land management. Traditional field-based biomass measurements, however, are labor-intensive and limited in spatial coverage. In this study, we estimated aboveground biomass of green vegetation across Mongolia using moderate-resolution remote sensing data and machine learning methods to support large-scale ecological assessment. Three spectral bands and nine vegetation indices derived from Moderate Resolution Imaging Spectroradiometer MOD13A1 data from August 2021 were combined with field measurements from 275 monitoring sites. Random forest, artificial neural network, and extreme gradient boosting \ models were trained and optimized using grid-search hyperparameter tuning and evaluated through cross-validation. Among the tested models, Random forest achieved the highest predictive accuracy (R² = 0.88, MAE = 68.45 kg ha⁻¹, RMSE = 109.93 kg ha⁻¹), followed by extreme gradient boosting (R² = 0.80, MAE = 87.47 kg ha⁻¹, RMSE = 143.09 kg ha⁻¹), while artificial neural network showed comparatively lower performance. The resulting biomass maps reveal pronounced spatial variability linked to Mongolia’s climatic and ecological gradients. Our findings demonstrate that machine learning methods combined with remote sensing provide an efficient framework for estimating vegetation biomass across Mongolia under the environmental conditions represented by the August 2021 dataset. The generated biomass products offer valuable ecological insights and can support rangeland monitoring, conservation planning, and assessments of ecosystem vulnerability under ongoing climate change.
- 6 April 2026
- Published by Mongolian Journals Online in Mongolian Geoscientist
- Vol. 31 (62) , 99-114
Open AccessGreen vegetation biomass estimation in Mongolia using RS data and machine-learning methodsing methods
Green vegetation biomass is a key ecological indicator reflecting ecosystem productivity, rangeland condition, and responses to climate variability, particularly in arid and semi-arid environments. In Mongolia, where vast grassland and desert-steppe ecosystems support both biodiversity and pastoral livelihoods, spatially explicit biomass information is essential for ecological monitoring and conservation-oriented land management. Traditional field-based biomass measurements, however, are labor-intensive and limited in spatial coverage. In this study, we estimated aboveground biomass of green vegetation across Mongolia using moderate-resolution remote sensing data and machine learning methods to support large-scale ecological assessment. Three spectral bands and nine vegetation indices derived from Moderate Resolution Imaging Spectroradiometer MOD13A1 data from August 2021 were combined with field measurements from 275 monitoring sites. Random forest, artificial neural network, and extreme gradient boosting \ models were trained and optimized using grid-search hyperparameter tuning and evaluated through cross-validation. Among the tested models, Random forest achieved the highest predictive accuracy (R² = 0.88, MAE = 68.45 kg ha⁻¹, RMSE = 109.93 kg ha⁻¹), followed by extreme gradient boosting (R² = 0.80, MAE = 87.47 kg ha⁻¹, RMSE = 143.09 kg ha⁻¹), while artificial neural network showed comparatively lower performance. The resulting biomass maps reveal pronounced spatial variability linked to Mongolia’s climatic and ecological gradients. Our findings demonstrate that machine learning methods combined with remote sensing provide an efficient framework for estimating vegetation biomass across Mongolia under the environmental conditions represented by the August 2021 dataset. The generated biomass products offer valuable ecological insights and can support rangeland monitoring, conservation planning, and assessments of ecosystem vulnerability under ongoing climate change.
- Open Access
- 30 March 2026
- Published by MDPI AG in Remote Sensing
- Vol. 18 (7) , 1036
SERF-XCH4: A Stacked Ensemble Framework for Spatiotemporal Continuous Methane Monitoring and Driver Analysis
Satellite observations of methane are frequently compromised by extensive data gaps caused by cloud cover and aerosol contamination, limiting their utility for continuous regional monitoring. To reconstruct these spatiotemporal discontinuities, this study developed the Stacked Ensemble Reconstruction Framework for Methane (SERF-XCH4). By integrating Sentinel-5P TROPOMI retrievals with 25 multi-source environmental covariates, we generated a spatiotemporally continuous, high-resolution (0.1°) monthly dataset (SERF-XCH4-IM) for Inner Mongolia spanning 2019 to 2023. Comprehensive validation demonstrates that the framework achieves exceptional predictive fidelity with a Coefficient of Determination () of 0.93 and a Root Mean Square Error (RMSE) of 7.89 ppb, significantly surpassing the performance of individual base learners and traditional interpolation methods. Furthermore, spatial block cross-validation confirmed robust generalization capabilities () in data-void regions. To unravel the “black box” of the model, SHapley Additive exPlanations (SHAP) analysis was employed, revealing that temporal factors (contributing 63.9%), air temperature, and elevation are the dominant drivers governing XCH4 variability. Spatiotemporal analysis further identified the Hulunbuir region as a significant growth “hotspot” with an annual increase rate exceeding 18.5 ppb/yr, a trend primarily driven by intensified emissions during the autumn and winter seasons. Consequently, this framework establishes a high-precision, interpretable paradigm for regional methane monitoring and geo-information reconstruction.- 30 March 2026
- Published by MDPI AG in Remote Sensing
- Vol. 18 (7) , 1036
Open AccessSERF-XCH4: A Stacked Ensemble Framework for Spatiotemporal Continuous Methane Monitoring and Driver Analysis
Satellite observations of methane are frequently compromised by extensive data gaps caused by cloud cover and aerosol contamination, limiting their utility for continuous regional monitoring. To reconstruct these spatiotemporal discontinuities, this study developed the Stacked Ensemble Reconstruction Framework for Methane (SERF-XCH4). By integrating Sentinel-5P TROPOMI retrievals with 25 multi-source environmental covariates, we generated a spatiotemporally continuous, high-resolution (0.1°) monthly dataset (SERF-XCH4-IM) for Inner Mongolia spanning 2019 to 2023. Comprehensive validation demonstrates that the framework achieves exceptional predictive fidelity with a Coefficient of Determination () of 0.93 and a Root Mean Square Error (RMSE) of 7.89 ppb, significantly surpassing the performance of individual base learners and traditional interpolation methods. Furthermore, spatial block cross-validation confirmed robust generalization capabilities () in data-void regions. To unravel the “black box” of the model, SHapley Additive exPlanations (SHAP) analysis was employed, revealing that temporal factors (contributing 63.9%), air temperature, and elevation are the dominant drivers governing XCH4 variability. Spatiotemporal analysis further identified the Hulunbuir region as a significant growth “hotspot” with an annual increase rate exceeding 18.5 ppb/yr, a trend primarily driven by intensified emissions during the autumn and winter seasons. Consequently, this framework establishes a high-precision, interpretable paradigm for regional methane monitoring and geo-information reconstruction. - Open Access
- 19 December 2025
- Published by Mongolian Journals Online in Mongolian Journal of Geography and Geoecology
- Vol. 62 (46) , 157-166
Assessing the Spatial Distribution of Urban Heat Island in Erdenet City Using Machine Learning Approaches
The Urban Heat Island (UHI) phenomenon represents a critical ecological challenge associated with rapid global urbanization, exerting adverse impacts on energy consumption, air quality, and public health. As Mongolia's major mining city, Erdenet is experiencing increasing environmental pressures during its urbanization process, necessitating comprehensive investigation of this phenomenon. This study aimed to analyze the spatiotemporal variations of UHI effects in Erdenet city over a 35-year period (1989-2024) and develop predictive models for future projections. Utilizing multi-temporal Landsat satellite imagery for spatially modeling Land Surface Temperature (LST), advanced machine learning algorithms including Random Forest (RF) and Extreme Gradient Boosting (XGBoost) were employed. The results revealed that mean LST increased from 23.05°C in 1989 to 31.15°C in 2004, subsequently decreasing to 25.70°C in 2024. A notable finding was the spatial redistribution of UHI patterns: during the initial years, elevated LST values were concentrated in areas with sparse vegetation cover and barren land, while by 2024, high temperatures had shifted to impervious surfaces including built-up areas and residential settlements. This spatial transition indicates an intensification of UHI effects driven by urban expansion. Throughout the study period, forest area decreased significantly, while water bodies and green infrastructure consistently maintained lower temperatures, demonstrating the crucial temperature-regulating role of these natural elements. These findings align with international research emphasizing the vital role of vegetation in mitigating UHI intensity, and highlight the imperative to integrate green infrastructure into urban planning to enhance climate resilience in rapidly urbanizing regions. Эрдэнэт хотын “Хотын дулааны арал”-ын орон зайн тархалтыг машин сургалтын аргаар үнэлэх нь ХУРААНГУЙ: “Хотын дулаан арал”-ийн (Urban Heat Island, UHI) үзэгдэл нь дэлхийн хурдацтай хотжилттой холбоотой гол экологийн асуудлын нэг бөгөөд эрчим хүчний хэрэглээ, агаарын чанар болон иргэдийн эрүүл мэндэд сөрөг нөлөөг үзүүлж байна. Монгол Улсын томоохон уул уурхайн хот Эрдэнэт нь хотжих явцдаа ийм төрлийн экологийн дарамт үүсэж байгаа тул энэ үзэгдлийг судлах шаардлага зайлшгүй гарч ирлээ. Энэхүү судалгаанд Эрдэнэт хотын UHI үзэгдлийн орон зайн өөрчлөлтийг 35 жилийн хугацаанд (1989-2024) шинжилж, цаашдын урьдчилсан таамаглал гаргах зорилго тавьсан. Олон цаг хугацааны Landsat хиймэл дагуулын мэдээллийг ашиглан газрын гадаргын температур (LST)-ийг орон зайд загварчлахдаа ахисан түвшний машин сургалтын Random Forest (RF) болон Extreme Gradient Boosting (XGBoost) алгоритмуудыг хэрэглэлээ. Судалгааны үр дүнгээс харахад LST-ийн дундаж утга 1989 онд 23.05°C байсан бол 2004 онд 31.15°C хүртэл өссөн ч 2024 онд 25.70°C болж буурчээ. Онцлох зүйл нь UHI-ийн орон зайн шилжилт юм: эхний жилүүдэд LST-ийн өндөр утга нь ургамлын бүрхэвч сийрэг, халцгай газарт төвлөрч байсан бол 2024 онд барилга байгууламж, гэр хороолол зэрэг ус үл нэвтрэх гадаргуу бүхий газруудад тэлжээ. Энэ нь хотын тэлэлтээс шалтгаалан UHI-ийн эрчим нэмэгдэж буйг илтгэнэ. Судалгааны хугацаанд ойн сангийн талбай эрс багассан харин усны сан, ногоон байгууламж бүхий газрууд тогтмол бага температуртай байсан нь эдгээр байгалийн элементүүдийн температур зохицуулах чухал үүргийг харуулж байна. Эдгээр үр дүн нь UHI-ийн эрчмийг бууруулахад ургамлын чухал үүргийг онцолсон олон улсын судалгаатай нийцэж, хурдацтай хотжиж буй бүс нутагт уур амьсгалын тэсвэртэй байдлыг нэмэгдүүлэх зорилгоор ногоон дэд бүтцийг хот төлөвлөлтөд нэгтгэх шаардлагыг харуулж байна. Түлхүүр үгс: Хотын дулааны арал, Машин сургалт, Газрын гадаргын температур
- 19 December 2025
- Published by Mongolian Journals Online in Mongolian Journal of Geography and Geoecology
- Vol. 62 (46) , 157-166
Open AccessAssessing the Spatial Distribution of Urban Heat Island in Erdenet City Using Machine Learning Approaches
The Urban Heat Island (UHI) phenomenon represents a critical ecological challenge associated with rapid global urbanization, exerting adverse impacts on energy consumption, air quality, and public health. As Mongolia's major mining city, Erdenet is experiencing increasing environmental pressures during its urbanization process, necessitating comprehensive investigation of this phenomenon. This study aimed to analyze the spatiotemporal variations of UHI effects in Erdenet city over a 35-year period (1989-2024) and develop predictive models for future projections. Utilizing multi-temporal Landsat satellite imagery for spatially modeling Land Surface Temperature (LST), advanced machine learning algorithms including Random Forest (RF) and Extreme Gradient Boosting (XGBoost) were employed. The results revealed that mean LST increased from 23.05°C in 1989 to 31.15°C in 2004, subsequently decreasing to 25.70°C in 2024. A notable finding was the spatial redistribution of UHI patterns: during the initial years, elevated LST values were concentrated in areas with sparse vegetation cover and barren land, while by 2024, high temperatures had shifted to impervious surfaces including built-up areas and residential settlements. This spatial transition indicates an intensification of UHI effects driven by urban expansion. Throughout the study period, forest area decreased significantly, while water bodies and green infrastructure consistently maintained lower temperatures, demonstrating the crucial temperature-regulating role of these natural elements. These findings align with international research emphasizing the vital role of vegetation in mitigating UHI intensity, and highlight the imperative to integrate green infrastructure into urban planning to enhance climate resilience in rapidly urbanizing regions. Эрдэнэт хотын “Хотын дулааны арал”-ын орон зайн тархалтыг машин сургалтын аргаар үнэлэх нь ХУРААНГУЙ: “Хотын дулаан арал”-ийн (Urban Heat Island, UHI) үзэгдэл нь дэлхийн хурдацтай хотжилттой холбоотой гол экологийн асуудлын нэг бөгөөд эрчим хүчний хэрэглээ, агаарын чанар болон иргэдийн эрүүл мэндэд сөрөг нөлөөг үзүүлж байна. Монгол Улсын томоохон уул уурхайн хот Эрдэнэт нь хотжих явцдаа ийм төрлийн экологийн дарамт үүсэж байгаа тул энэ үзэгдлийг судлах шаардлага зайлшгүй гарч ирлээ. Энэхүү судалгаанд Эрдэнэт хотын UHI үзэгдлийн орон зайн өөрчлөлтийг 35 жилийн хугацаанд (1989-2024) шинжилж, цаашдын урьдчилсан таамаглал гаргах зорилго тавьсан. Олон цаг хугацааны Landsat хиймэл дагуулын мэдээллийг ашиглан газрын гадаргын температур (LST)-ийг орон зайд загварчлахдаа ахисан түвшний машин сургалтын Random Forest (RF) болон Extreme Gradient Boosting (XGBoost) алгоритмуудыг хэрэглэлээ. Судалгааны үр дүнгээс харахад LST-ийн дундаж утга 1989 онд 23.05°C байсан бол 2004 онд 31.15°C хүртэл өссөн ч 2024 онд 25.70°C болж буурчээ. Онцлох зүйл нь UHI-ийн орон зайн шилжилт юм: эхний жилүүдэд LST-ийн өндөр утга нь ургамлын бүрхэвч сийрэг, халцгай газарт төвлөрч байсан бол 2024 онд барилга байгууламж, гэр хороолол зэрэг ус үл нэвтрэх гадаргуу бүхий газруудад тэлжээ. Энэ нь хотын тэлэлтээс шалтгаалан UHI-ийн эрчим нэмэгдэж буйг илтгэнэ. Судалгааны хугацаанд ойн сангийн талбай эрс багассан харин усны сан, ногоон байгууламж бүхий газрууд тогтмол бага температуртай байсан нь эдгээр байгалийн элементүүдийн температур зохицуулах чухал үүргийг харуулж байна. Эдгээр үр дүн нь UHI-ийн эрчмийг бууруулахад ургамлын чухал үүргийг онцолсон олон улсын судалгаатай нийцэж, хурдацтай хотжиж буй бүс нутагт уур амьсгалын тэсвэртэй байдлыг нэмэгдүүлэх зорилгоор ногоон дэд бүтцийг хот төлөвлөлтөд нэгтгэх шаардлагыг харуулж байна. Түлхүүр үгс: Хотын дулааны арал, Машин сургалт, Газрын гадаргын температур
- Open Access
- 19 December 2025
- Published by Mongolian Journals Online in Mongolian Journal of Geography and Geoecology
- Vol. 62 (46) , 101-112
Modeling the spatial distribution of carbon dioxide concentration by machine learning
The rapid increase in atmospheric carbon dioxide (CO₂) is a significant contributor to modern climate change and global warming. High-resolution data are essential for scientifically valid studies of CO₂ dynamics and changes. Although greenhouse gas observation satellites exist, they are still inadequate in terms of spatiotemporal resolution. We modeled to obtain a high-spatiotemporal resolution (16-day, 0.1x0.1) XCO₂ by machine learning algorithms (such as Random Forest (RF), eXtreme Gradient Boosting (XGB), Light Gradient Boosting Machine (LGBM), and Stacking (Stk)) using multisource data (such as weather parameters, vegetation index, land cover, elevation and population density) over Mongolia and China from 2010 to 2022. The results indicated that the XGB demonstrated good performance, with R² values exceeding 0.95 and 10-fold cross-validation R² (cv) values exceeding 0.55 during the study periods. Additionally, when calculating the Shapley Additive Explanations (SHAP) values based on the XGB model's results, the weather variables (53%) had the largest influence on the distribution of XCO₂. In comparison, population density alone had an 11% influence. As a result, the average XGB XCO₂ over the study period was 402.3±1.14 ppm, with the highest value observed in eastern China and the lowest value observed in Mongolia and the Qinghai-Tibet Plateau. This is related to the region's economic development and human activities. In terms of seasonal dynamics, the XGB XCO₂ values for winter, spring, summer, and autumn were 403.4±1.2 ppm, 403.3±1.3 ppm, 400.9±1.5 ppm, and 402.1±1.0 ppm, respectively. This shows that XCO₂ is highest in winter and lowest in summer. This seasonal dynamic is because plants absorb more CO₂ through photosynthesis in summer than in other seasons. The atmospheric ХCO₂ map produced in this study can serve as a baseline map for studying the carbon cycle in terrestrial ecosystems and climate change. Нүүрсхүчлийн хийн агууламжийн орон зайн тархалтыг машин сургалтаар загварчлах нь ХУРААНГУЙ: Агаар мандал дахь нүүрсхүчлийн хий (CO₂)-н агууламжийн хурдацтай өсөлт нь Дэлхийн дулаарал, орчин үеийн уур амьсгалын өөрчлөлтийн үндсэн шалтгаан болж байна. CO-н динамик, өөрчлөлтийн талаарх өндөр нарийвчлалтай өгөгдөлд тулгуурласан судалгаа хийх нь чухал юм. Хэдийгээр хүлэмжийн хийн ажиглалтын хиймэл дагуулууд байгаа боловч цаг хугацаа, орон зайн хувьд шаардлага хангахгүй хэвээр байна. Иймд энэхүү судалгаандаа, бид Монгол болон БНХАУ (Бүгд Найрамдах Хятад Ард Улс)-ын нутаг дэвсгэрийн хэмжээнд, 2010-аас 2022 он хүртэлх хугацааны Хүлэмжийн хийн ажиглалтын хиймэл дагуул (GOSAT)-ын хуурай агаарын босоо баганын дундаж ХCO₂ болон цаг уурын хэмжигдэхүүнүүд, ургамлын индекс, газрын бүрхэвч, гадаргын өндөршил, хүн амын нягтшил зэрэг 12 өгөгдлийг ашиглан машин сургалтын 4 өөр аргаар (Random Forest (RF), eXtreme Gradient Boosting (XGB), Light Gradient Boosting Machine (LGBM), Stacking (Stk)) XCO₂-г загварчлан харьцуулсан. Эдгээрээс XGB XCO₂ загвар хамгийн өндөр гүйцэтгэлтэй буюу детерминацийн коэффициент R²>0.95, солбин бататгах R²(cv)>0.55 байлаа. Түүнчлэн XGB загварын үр дүнд тулгуурлан Таамаглалыг тайлбарлах нэгдсэн тогтолцоо буюу Shapley Additive Explanations (SHAP) утгыг тооцоход XCO₂-н тархалтад цаг уурын хувьсагчид хамгийн их нөлөө (53%) үзүүлсэн бол хүн амын нягтшил дангаараа 11% нөлөө үзүүлсэн. Өндөр гүйцэтгэлтэй XGB загвараар судалгааны талбайг хамарсан өндөр нарийвчлалтай (16 өдрийн, 0.1°x0.1°) XGB XCO₂-н орон зайн тасралтгүй өгөгдлийг гарган авсан. Үр дүнд судалгааны хугацааны талбайн дундаж XGB XCO₂ нь 402.3±1.14 ppm, их утга нь БНХАУ-ын зүүн хэсгээр, бага утга нь Монгол болон Чинхай-Төвөдийн өндөрлөгт ажиглагдсан. Энэ нь бүс нутгийн эдийн засгийн хөгжил, хүний үйл ажиллагаатай холбоотой юм. Улирлын динамикийн хувьд XGB XCO₂ утга нь өвөл, хавар, зун, намрын улиралд харгалзан 403.4±1.2 ppm, 403.3±1.3 ppm, 400.9±1.5 ppm, 402.1±1.0 ppm байлаа. Эндээс үзвэл XCO₂ нь өвлийн улиралд хамгийн их утгатай, зун хамгийн бага утгатай байна. Улирлын динамик нь зуны улиралд ургамал фотосинтезийн процессоор СО-г бусад улиралтай харьцуулахад ихээр шингээдэгтэй холбоотой. Энэхүү судалгааны ажлын үр дүн нь биосферийн нүүрстөрөгчийн эргэлтийг загварчлах, уур амьсгалын өөрчлөлтийг судлахад суурь өгөгдөл болно....
- 19 December 2025
- Published by Mongolian Journals Online in Mongolian Journal of Geography and Geoecology
- Vol. 62 (46) , 101-112
Open AccessModeling the spatial distribution of carbon dioxide concentration by machine learning
The rapid increase in atmospheric carbon dioxide (CO₂) is a significant contributor to modern climate change and global warming. High-resolution data are essential for scientifically valid studies of CO₂ dynamics and changes. Although greenhouse gas observation satellites exist, they are still inadequate in terms of spatiotemporal resolution. We modeled to obtain a high-spatiotemporal resolution (16-day, 0.1x0.1) XCO₂ by machine learning algorithms (such as Random Forest (RF), eXtreme Gradient Boosting (XGB), Light Gradient Boosting Machine (LGBM), and Stacking (Stk)) using multisource data (such as weather parameters, vegetation index, land cover, elevation and population density) over Mongolia and China from 2010 to 2022. The results indicated that the XGB demonstrated good performance, with R² values exceeding 0.95 and 10-fold cross-validation R² (cv) values exceeding 0.55 during the study periods. Additionally, when calculating the Shapley Additive Explanations (SHAP) values based on the XGB model's results, the weather variables (53%) had the largest influence on the distribution of XCO₂. In comparison, population density alone had an 11% influence. As a result, the average XGB XCO₂ over the study period was 402.3±1.14 ppm, with the highest value observed in eastern China and the lowest value observed in Mongolia and the Qinghai-Tibet Plateau. This is related to the region's economic development and human activities. In terms of seasonal dynamics, the XGB XCO₂ values for winter, spring, summer, and autumn were 403.4±1.2 ppm, 403.3±1.3 ppm, 400.9±1.5 ppm, and 402.1±1.0 ppm, respectively. This shows that XCO₂ is highest in winter and lowest in summer. This seasonal dynamic is because plants absorb more CO₂ through photosynthesis in summer than in other seasons. The atmospheric ХCO₂ map produced in this study can serve as a baseline map for studying the carbon cycle in terrestrial ecosystems and climate change. Нүүрсхүчлийн хийн агууламжийн орон зайн тархалтыг машин сургалтаар загварчлах нь ХУРААНГУЙ: Агаар мандал дахь нүүрсхүчлийн хий (CO₂)-н агууламжийн хурдацтай өсөлт нь Дэлхийн дулаарал, орчин үеийн уур амьсгалын өөрчлөлтийн үндсэн шалтгаан болж байна. CO-н динамик, өөрчлөлтийн талаарх өндөр нарийвчлалтай өгөгдөлд тулгуурласан судалгаа хийх нь чухал юм. Хэдийгээр хүлэмжийн хийн ажиглалтын хиймэл дагуулууд байгаа боловч цаг хугацаа, орон зайн хувьд шаардлага хангахгүй хэвээр байна. Иймд энэхүү судалгаандаа, бид Монгол болон БНХАУ (Бүгд Найрамдах Хятад Ард Улс)-ын нутаг дэвсгэрийн хэмжээнд, 2010-аас 2022 он хүртэлх хугацааны Хүлэмжийн хийн ажиглалтын хиймэл дагуул (GOSAT)-ын хуурай агаарын босоо баганын дундаж ХCO₂ болон цаг уурын хэмжигдэхүүнүүд, ургамлын индекс, газрын бүрхэвч, гадаргын өндөршил, хүн амын нягтшил зэрэг 12 өгөгдлийг ашиглан машин сургалтын 4 өөр аргаар (Random Forest (RF), eXtreme Gradient Boosting (XGB), Light Gradient Boosting Machine (LGBM), Stacking (Stk)) XCO₂-г загварчлан харьцуулсан. Эдгээрээс XGB XCO₂ загвар хамгийн өндөр гүйцэтгэлтэй буюу детерминацийн коэффициент R²>0.95, солбин бататгах R²(cv)>0.55 байлаа. Түүнчлэн XGB загварын үр дүнд тулгуурлан Таамаглалыг тайлбарлах нэгдсэн тогтолцоо буюу Shapley Additive Explanations (SHAP) утгыг тооцоход XCO₂-н тархалтад цаг уурын хувьсагчид хамгийн их нөлөө (53%) үзүүлсэн бол хүн амын нягтшил дангаараа 11% нөлөө үзүүлсэн. Өндөр гүйцэтгэлтэй XGB загвараар судалгааны талбайг хамарсан өндөр нарийвчлалтай (16 өдрийн, 0.1°x0.1°) XGB XCO₂-н орон зайн тасралтгүй өгөгдлийг гарган авсан. Үр дүнд судалгааны хугацааны талбайн дундаж XGB XCO₂ нь 402.3±1.14 ppm, их утга нь БНХАУ-ын зүүн хэсгээр, бага утга нь Монгол болон Чинхай-Төвөдийн өндөрлөгт ажиглагдсан. Энэ нь бүс нутгийн эдийн засгийн хөгжил, хүний үйл ажиллагаатай холбоотой юм. Улирлын динамикийн хувьд XGB XCO₂ утга нь өвөл, хавар, зун, намрын улиралд харгалзан 403.4±1.2 ppm, 403.3±1.3 ppm, 400.9±1.5 ppm, 402.1±1.0 ppm байлаа. Эндээс үзвэл XCO₂ нь өвлийн улиралд хамгийн их утгатай, зун хамгийн бага утгатай байна. Улирлын динамик нь зуны улиралд ургамал фотосинтезийн процессоор СО-г бусад улиралтай харьцуулахад ихээр шингээдэгтэй холбоотой. Энэхүү судалгааны ажлын үр дүн нь биосферийн нүүрстөрөгчийн эргэлтийг загварчлах, уур амьсгалын өөрчлөлтийг судлахад суурь өгөгдөл болно....
- Open Access
- 19 December 2025
- Published by Mongolian Journals Online in Mongolian Journal of Geography and Geoecology
- Vol. 62 (46) , 119-128
A Study on the Methodological Approaches for Estimating Pasture Biomass Using Multi-Source Remote Sensing Data
In recent years, remote sensing technologies, particularly the use of multi-source datasets with high spectral and spatial resolution, have become increasingly important for estimating in above ground biomass (AGB). The integration of artificial intelligence and machine learning methods enables more accurate and efficient biomass modeling. This study aims to estimate and map grassland biomass in Bayandelger soum using field-measured biomass data and high-resolution Sentinel-2 satellite imagery through five machine learning algorithms: Random Forest Regression (RF), Support Vector Regression (SVR), Extra Trees Regression (ETR), Gradient Boosted Regression Trees (GBRT), and Extreme Gradient Boosting (XGB). Field measurements were conducted at 65 sampling sites in June 2020. The results showed that the average above-ground biomass was 37.9 g/m² in Carex pp-dominated pastures (22 points), 30.1 g/m² in Stipa sp -dominated pastures (24 points), 64.9 g/m² in Artemisia sp-dominated pastures (8 points), 22.4 g/m² in Cleistogenes squarrosa-dominated pastures (5 points), and 34.4 g/m² in Agropyron repens (L.) P. B -dominated pastures (3 points). Among the five models, the comparison between estimated and field-measured biomass indicated that SVR achieved the highest performance (RMSE = 67.1 kg/ha, R² = 0.70), followed by GBRT (RMSE = 72.1 kg/ha, R² = 0.65), RF (RMSE = 72.3 kg/ha, R² = 0.65), ETR (RMSE = 72.5 kg/ha, R² = 0.65), and XGB (RMSE = 83.9 kg/ha, R² = 0.53). Our findings demonstrate that machine learning models, combined with Sentinel-2 imagery, can effectively estimate and map grassland biomass at the regional scale. This research provides valuable insights for assessing grassland degradation, desertification, and improving sustainable rangeland management in Mongolia. Зайнаас тандан судлалын олон эх сурвалжийн өгөгдлийг ашиглан бэлчээрийн биомассыг тооцоолох арга зүйн асуудалд (Төв аймгийн Баяндэлгэр сумын жишээн дээр) ХУРААНГУЙ: Орчин үед бэлчээрийн биомассын судалгаанд зайнаас тандан судлал тэр дундаа спектрийн өндөр нарийвчлал бүхий олон эх сурвалжийн өгөгдөл, ялангуяа хиймэл оюун ухааны арга зүйг ашиглах нь нэн чухал юм. Энэхүү судалгааны ажлын зорилго нь хээрийн хэмжилтээр бэлчээрийн ургамлын биомассыг тодорхойлж, өндөр нарийвчлал бүхий Sentinel-2 хиймэл дагуулын өгөгдөл ашиглан хиймэл оюуны машин сургалтын 5 загвар (санамсаргүй ойн регрессийн алгоритм /Random forest regression algorithm/, тулах векторын регрессийн алгоритм/Support Vector Regression algorithm, сайжруулсан модны регрессийн алгоритм/Extra tree regression algorithm/, градиентаар хөгжих/Gradient Boosted Regression tree, хүчтэй градиентаар хөгжих/Extreme Gradient Boosting)-аар Баяндэлгэр сумын хэмжээнд бэлчээрийн биомассыг зураглахыг зорилоо. Судалгааны талбайн хэмжээнд 2020 оны 6 – р сард бэлчээрийн ургамлын биомассыг нийт 65 цэгт хэмжилт хийлээ. Бэлчээрийн биомассын боловсруулсан үр дүнгээс үзвэл улалж зонхилсон бэлчээрийн биомассын дундаж хэмжээ 37.9 гр/м² (22 цэгт), хялгана зонхилсон бэлчээрт 30.1 гр/м² (24 цэгт), шарилж зонхилсон бэлчээрт 64.9 гр/м² (8 цэгт), хазаар өвс зонхилсон бэлчээрт 22.4 гр/м² (5 цэгт), хиаг зонхилсон бэлчээрт 34.4 гр/м² (3 цэгт) тус тус байлаа. Баяндэлгэр сумын хэмжээнд бэлчээрийн биомассын машин сургалтын 5 загвараар тооцоолсон үр дүн болон хээрийн хэмжилтийн хооронд өндөр хамааралтай RF дундаж квадрат алдаа (RMSE)=72.3 кг/га, детерминацийн коэффициент (R²)=0.65, XGB (RMSE=83.9 кг/га, R²=0.53), GBRT (RMSE=72.1 кг/га, R²=0.65), ETR (RMSE=72.5 кг/га, R²=0.65) байсан бол хамгийн өндөр гүйцэтгэлтэй загвар нь SVR (RMSE=67.1 кг/га, R²=0.7) байлаа. Бидний судалгааны ажил нь бэлчээрийн доройтол, талхагдал, менежментэд онолын болон практик ач холбогдолтой юм. Түлхүүр үг: Машин сургалт, Sentinel, Биомасс
- 19 December 2025
- Published by Mongolian Journals Online in Mongolian Journal of Geography and Geoecology
- Vol. 62 (46) , 119-128
Open AccessA Study on the Methodological Approaches for Estimating Pasture Biomass Using Multi-Source Remote Sensing Data
In recent years, remote sensing technologies, particularly the use of multi-source datasets with high spectral and spatial resolution, have become increasingly important for estimating in above ground biomass (AGB). The integration of artificial intelligence and machine learning methods enables more accurate and efficient biomass modeling. This study aims to estimate and map grassland biomass in Bayandelger soum using field-measured biomass data and high-resolution Sentinel-2 satellite imagery through five machine learning algorithms: Random Forest Regression (RF), Support Vector Regression (SVR), Extra Trees Regression (ETR), Gradient Boosted Regression Trees (GBRT), and Extreme Gradient Boosting (XGB). Field measurements were conducted at 65 sampling sites in June 2020. The results showed that the average above-ground biomass was 37.9 g/m² in Carex pp-dominated pastures (22 points), 30.1 g/m² in Stipa sp -dominated pastures (24 points), 64.9 g/m² in Artemisia sp-dominated pastures (8 points), 22.4 g/m² in Cleistogenes squarrosa-dominated pastures (5 points), and 34.4 g/m² in Agropyron repens (L.) P. B -dominated pastures (3 points). Among the five models, the comparison between estimated and field-measured biomass indicated that SVR achieved the highest performance (RMSE = 67.1 kg/ha, R² = 0.70), followed by GBRT (RMSE = 72.1 kg/ha, R² = 0.65), RF (RMSE = 72.3 kg/ha, R² = 0.65), ETR (RMSE = 72.5 kg/ha, R² = 0.65), and XGB (RMSE = 83.9 kg/ha, R² = 0.53). Our findings demonstrate that machine learning models, combined with Sentinel-2 imagery, can effectively estimate and map grassland biomass at the regional scale. This research provides valuable insights for assessing grassland degradation, desertification, and improving sustainable rangeland management in Mongolia. Зайнаас тандан судлалын олон эх сурвалжийн өгөгдлийг ашиглан бэлчээрийн биомассыг тооцоолох арга зүйн асуудалд (Төв аймгийн Баяндэлгэр сумын жишээн дээр) ХУРААНГУЙ: Орчин үед бэлчээрийн биомассын судалгаанд зайнаас тандан судлал тэр дундаа спектрийн өндөр нарийвчлал бүхий олон эх сурвалжийн өгөгдөл, ялангуяа хиймэл оюун ухааны арга зүйг ашиглах нь нэн чухал юм. Энэхүү судалгааны ажлын зорилго нь хээрийн хэмжилтээр бэлчээрийн ургамлын биомассыг тодорхойлж, өндөр нарийвчлал бүхий Sentinel-2 хиймэл дагуулын өгөгдөл ашиглан хиймэл оюуны машин сургалтын 5 загвар (санамсаргүй ойн регрессийн алгоритм /Random forest regression algorithm/, тулах векторын регрессийн алгоритм/Support Vector Regression algorithm, сайжруулсан модны регрессийн алгоритм/Extra tree regression algorithm/, градиентаар хөгжих/Gradient Boosted Regression tree, хүчтэй градиентаар хөгжих/Extreme Gradient Boosting)-аар Баяндэлгэр сумын хэмжээнд бэлчээрийн биомассыг зураглахыг зорилоо. Судалгааны талбайн хэмжээнд 2020 оны 6 – р сард бэлчээрийн ургамлын биомассыг нийт 65 цэгт хэмжилт хийлээ. Бэлчээрийн биомассын боловсруулсан үр дүнгээс үзвэл улалж зонхилсон бэлчээрийн биомассын дундаж хэмжээ 37.9 гр/м² (22 цэгт), хялгана зонхилсон бэлчээрт 30.1 гр/м² (24 цэгт), шарилж зонхилсон бэлчээрт 64.9 гр/м² (8 цэгт), хазаар өвс зонхилсон бэлчээрт 22.4 гр/м² (5 цэгт), хиаг зонхилсон бэлчээрт 34.4 гр/м² (3 цэгт) тус тус байлаа. Баяндэлгэр сумын хэмжээнд бэлчээрийн биомассын машин сургалтын 5 загвараар тооцоолсон үр дүн болон хээрийн хэмжилтийн хооронд өндөр хамааралтай RF дундаж квадрат алдаа (RMSE)=72.3 кг/га, детерминацийн коэффициент (R²)=0.65, XGB (RMSE=83.9 кг/га, R²=0.53), GBRT (RMSE=72.1 кг/га, R²=0.65), ETR (RMSE=72.5 кг/га, R²=0.65) байсан бол хамгийн өндөр гүйцэтгэлтэй загвар нь SVR (RMSE=67.1 кг/га, R²=0.7) байлаа. Бидний судалгааны ажил нь бэлчээрийн доройтол, талхагдал, менежментэд онолын болон практик ач холбогдолтой юм. Түлхүүр үг: Машин сургалт, Sentinel, Биомасс
- Open Access
- 1 September 2025
- Published by Mongolian Journals Online in Mongolian Journal of Geography and Geoecology
- Vol. 62 (46) , 8-14
Comparison of machine learning, supervised, and unsupervised algorithms for a land cover classification in northern Mongolia
The aim of this study is to compare the performance of machine learning approaches alongside supervised and unsupervised techniques to differentiate land cover classes in the northern Mongolia. The primary goal is to identify which of these methods can achieve the highest classification accuracy. For this analysis, we selected ten original spectral bands from Sentinel-2 data and utilized three different feature combinations. To differentiate among the available classes, we employed a support vector machine (SVM), a Mahalanobis distance classifier, and K-means clustering, assessing their relative effectiveness. In the three-band feature combination, K-means obtained the lowest accuracy at 70.08%, whereas SVM achieved 92.72%, ranking it as the most effective method. The Mahalanobis distance classifier closely followed with an accuracy of 90.36%. In the five-band combination, K-means improved its accuracy to 95.56%, surpassing earlier results, while the Mahalanobis distance achieved 95.07%, and SVM recorded an accuracy of 93.33%. In the analysis involving all ten bands, K-means again delivered the highest accuracy at 95.83%. The Mahalanobis distance classifier reached an accuracy of 93.93%, while SVM had an accuracy of 93.14%. In many cases, machine learning techniques can often outperform traditional methods. However, in this study, the traditional unsupervised technique surpassed both machine learning and supervised techniques in two cases. Thus, the results suggest that achieving high accuracy is not invariably attainable with machine learning or supervised image classification methods. In many instances, it depends on the selection of parameters, the data structure, and the radiometric properties of the objects of classes.
- 1 September 2025
- Published by Mongolian Journals Online in Mongolian Journal of Geography and Geoecology
- Vol. 62 (46) , 8-14
Open AccessComparison of machine learning, supervised, and unsupervised algorithms for a land cover classification in northern Mongolia
The aim of this study is to compare the performance of machine learning approaches alongside supervised and unsupervised techniques to differentiate land cover classes in the northern Mongolia. The primary goal is to identify which of these methods can achieve the highest classification accuracy. For this analysis, we selected ten original spectral bands from Sentinel-2 data and utilized three different feature combinations. To differentiate among the available classes, we employed a support vector machine (SVM), a Mahalanobis distance classifier, and K-means clustering, assessing their relative effectiveness. In the three-band feature combination, K-means obtained the lowest accuracy at 70.08%, whereas SVM achieved 92.72%, ranking it as the most effective method. The Mahalanobis distance classifier closely followed with an accuracy of 90.36%. In the five-band combination, K-means improved its accuracy to 95.56%, surpassing earlier results, while the Mahalanobis distance achieved 95.07%, and SVM recorded an accuracy of 93.33%. In the analysis involving all ten bands, K-means again delivered the highest accuracy at 95.83%. The Mahalanobis distance classifier reached an accuracy of 93.93%, while SVM had an accuracy of 93.14%. In many cases, machine learning techniques can often outperform traditional methods. However, in this study, the traditional unsupervised technique surpassed both machine learning and supervised techniques in two cases. Thus, the results suggest that achieving high accuracy is not invariably attainable with machine learning or supervised image classification methods. In many instances, it depends on the selection of parameters, the data structure, and the radiometric properties of the objects of classes.
- Open Access
- 1 September 2025
- Published by Mongolian Journals Online in Mongolian Journal of Geography and Geoecology
- Vol. 62 (46) , 132-138
Estimating the spatial distribution of urban heat islands in six central districts of Ulaanbaatar city using machine learning algorithm
Urban Heat Islands (UHIs) represent a growing environmental challenge in rapidly urbanizing cities, particularly in extreme climate regions such as Ulaanbaatar, Mongolia. This study aims to model and analyze the spatial distribution of UHIs across six central districts of Ulaanbaatar using Landsat 5 and Landsat 8 satellite imagery. Land Surface Temperature (LST) was derived from Landsat 8, and its relationship with six spectral indices the Normalized Difference Vegetation Index (NDVI), Normalized Difference Built-up Index (NDBI), Modified Normalized Difference Water Index (MNDWI), Bare Soil index (BI), Normalized Differences Water Index (NDWI), and Normalized Building Soil Index (NDBSI) was assessed. Two ensemble machine learning algorithms, Random Forest (RF) and Extreme Gradient Boosting (XGB), were employed to model UHI patterns. RF performed better in 1994 (a coefficients of determination (R²) = 0.72, a root-mean-square error (RMSE) = 1.89°C), while XGB showed superior performance in 2024 (R2 = 0.75, RMSE = 1.97°C). BI and NDBSI were the most influential contributors to UHI intensity, whereas vegetated areas (NDVI, MNDVI) had consistent cooling effects. The spatial modeling revealed a clear intensification of UHI effects over time, particularly in built-up and bare land zones. This study is significant as it applies ML techniques to long-term satellite data in a cold-climate, data-limited urban setting. By providing a rare longitudinal perspective, it contributes to understanding UHI dynamics in Mongolia and demonstrates the potential of open-access remote sensing data combined with ML for urban climate assessments. The findings offer valuable insight for urban planners, emphasizing the critical role of green infrastructure in mitigating thermal stress and informing climate-resilient development strategies in rapidly growing cities.
- 1 September 2025
- Published by Mongolian Journals Online in Mongolian Journal of Geography and Geoecology
- Vol. 62 (46) , 132-138
Open AccessEstimating the spatial distribution of urban heat islands in six central districts of Ulaanbaatar city using machine learning algorithm
Urban Heat Islands (UHIs) represent a growing environmental challenge in rapidly urbanizing cities, particularly in extreme climate regions such as Ulaanbaatar, Mongolia. This study aims to model and analyze the spatial distribution of UHIs across six central districts of Ulaanbaatar using Landsat 5 and Landsat 8 satellite imagery. Land Surface Temperature (LST) was derived from Landsat 8, and its relationship with six spectral indices the Normalized Difference Vegetation Index (NDVI), Normalized Difference Built-up Index (NDBI), Modified Normalized Difference Water Index (MNDWI), Bare Soil index (BI), Normalized Differences Water Index (NDWI), and Normalized Building Soil Index (NDBSI) was assessed. Two ensemble machine learning algorithms, Random Forest (RF) and Extreme Gradient Boosting (XGB), were employed to model UHI patterns. RF performed better in 1994 (a coefficients of determination (R²) = 0.72, a root-mean-square error (RMSE) = 1.89°C), while XGB showed superior performance in 2024 (R2 = 0.75, RMSE = 1.97°C). BI and NDBSI were the most influential contributors to UHI intensity, whereas vegetated areas (NDVI, MNDVI) had consistent cooling effects. The spatial modeling revealed a clear intensification of UHI effects over time, particularly in built-up and bare land zones. This study is significant as it applies ML techniques to long-term satellite data in a cold-climate, data-limited urban setting. By providing a rare longitudinal perspective, it contributes to understanding UHI dynamics in Mongolia and demonstrates the potential of open-access remote sensing data combined with ML for urban climate assessments. The findings offer valuable insight for urban planners, emphasizing the critical role of green infrastructure in mitigating thermal stress and informing climate-resilient development strategies in rapidly growing cities.
- Open Access
- 8 May 2025
- Published by MDPI AG in Atmosphere
- Vol. 16 (5) , 560
Spatiotemporal Characteristic of XCO2 and Its Changing Contribution Rate from Different Influencing Indicators in Mongolian Plateau of Central Asia
The Mongolian Plateau plays a crucial role in global carbon cycling, but the spatiotemporal characteristics of XCO2 concentration and its driving mechanism remain insufficiently explored. To solve this scientific issue, the synergistic methodology of mathematical statistics—the Pearson correlation and random forest model—was established using the main source of Orbiting Carbon Observatory 2 (OCO-2) satellite data. Results indicate the following: (1) Average XCO2 concentration of the Mongolian Plateau was 412 ppm, with an annual growth rate of 2.29 ppm/a from 2018 to 2022, along with higher values in the south and lower values in the north. The seasonal change displayed a clear temporal feature, in the order of spring (414.83 ppm) > winter (413.4 ppm) > autumn (411.3 ppm) > summer (409.12 ppm). The spatial distributions in spring, autumn, and winter were relatively consistent, all showing higher XCO2 concentrations in the east and lower concentrations in the west, whereas summer exhibited the opposite pattern. (2) From the perspective of the natural environment, XCO2 change was negatively correlated with the normalized difference vegetation index (NDVI), precipitation (PRE), and temperature (TEMP). Temporal analysis further revealed that this negative correlation was most pronounced in the eastern region, in which these three elements were all relatively high. (3) According to the random forest model, the influence of both single and interactive factors on the plateau’s XCO2 varied significantly. A comparison of driving factors revealed that the NDVI had the highest contribution rate (0.35), followed by fossil fuel combustion emissions (ODIAC), wind direction (WD), and wind speed (WS). As for interaction effects, the combination of NDVI and ODIAC showed the highest contribution rate (over 0.25), indicating a strong joint influence on XCO2. Other important interactions included WS and WD, ODIAC and WS, and NDVI and WS (all above 0.05). These findings provide valuable insights into the driving mechanisms of XCO2 on the Mongolian Plateau, offering a reference for regional carbon emission reduction policies.
- 8 May 2025
- Published by MDPI AG in Atmosphere
- Vol. 16 (5) , 560
Open AccessSpatiotemporal Characteristic of XCO2 and Its Changing Contribution Rate from Different Influencing Indicators in Mongolian Plateau of Central Asia
The Mongolian Plateau plays a crucial role in global carbon cycling, but the spatiotemporal characteristics of XCO2 concentration and its driving mechanism remain insufficiently explored. To solve this scientific issue, the synergistic methodology of mathematical statistics—the Pearson correlation and random forest model—was established using the main source of Orbiting Carbon Observatory 2 (OCO-2) satellite data. Results indicate the following: (1) Average XCO2 concentration of the Mongolian Plateau was 412 ppm, with an annual growth rate of 2.29 ppm/a from 2018 to 2022, along with higher values in the south and lower values in the north. The seasonal change displayed a clear temporal feature, in the order of spring (414.83 ppm) > winter (413.4 ppm) > autumn (411.3 ppm) > summer (409.12 ppm). The spatial distributions in spring, autumn, and winter were relatively consistent, all showing higher XCO2 concentrations in the east and lower concentrations in the west, whereas summer exhibited the opposite pattern. (2) From the perspective of the natural environment, XCO2 change was negatively correlated with the normalized difference vegetation index (NDVI), precipitation (PRE), and temperature (TEMP). Temporal analysis further revealed that this negative correlation was most pronounced in the eastern region, in which these three elements were all relatively high. (3) According to the random forest model, the influence of both single and interactive factors on the plateau’s XCO2 varied significantly. A comparison of driving factors revealed that the NDVI had the highest contribution rate (0.35), followed by fossil fuel combustion emissions (ODIAC), wind direction (WD), and wind speed (WS). As for interaction effects, the combination of NDVI and ODIAC showed the highest contribution rate (over 0.25), indicating a strong joint influence on XCO2. Other important interactions included WS and WD, ODIAC and WS, and NDVI and WS (all above 0.05). These findings provide valuable insights into the driving mechanisms of XCO2 on the Mongolian Plateau, offering a reference for regional carbon emission reduction policies.
- Open Access
- 11 March 2025
- Published by Springer Nature in Environmental Monitoring and Assessment
- Vol. 197 (4) , 1-13
Ecological risk evaluation of heavy metals based on hyperspectral: a case study of rice paddy soil in Xiangtan County, China
Heavy metal pollution in agricultural soils pose significant risks to food safety and human health, necessitating effective monitoring and assessment techniques. Recognizing characteristic spectral features is essential for accurately assessing the ecological risks associated with heavy metal pollution. To explore the potential of utilizing the hyperspectrum to evaluate ecologial risks of heavy metals pollution in paddy soil, the study investigated ecological risks and spectral characteristics (350 ~ 2500 nm) of heavy metals (Cd, Pb, Cr, Cu) in rice paddy soil in Xiangtan County, Hunan Province, China. Potential ecological risk index (RI) was employed to evaluate pollution level, and derivative transformations (first derivative R′ and second derivative R″) were applied to eliminate the hyperspectral noise. Results identified Cd as the primary pollutant in the paddy soil of Xiangtan County, followed by Pb, Cu, and Cr. Serious heavy metal pollution concentrated in riverine and cropland areas. After applying spectral transformations, the correlation between spectral reflectance and ecological risks significantly improved. The R″ transformation demonstrated the highest correlations for Cd (r = 0.95) at 2046 nm in high ecological risk category, while the R′ transformation exhibited the strongest negative correlation for Cd (r = − 0.94) at 2059 nm. This study highlights the potential of hyperspectral analysis as an effective tool for estimating ecological risks posed by heavy metal pollution in rice paddy soils. The findings provide valuable insights for improving monitoring practices and informing risk management strategies in agricultural environments.
- 11 March 2025
- Published by Springer Nature in Environmental Monitoring and Assessment
- Vol. 197 (4) , 1-13
Open AccessEcological risk evaluation of heavy metals based on hyperspectral: a case study of rice paddy soil in Xiangtan County, China
Heavy metal pollution in agricultural soils pose significant risks to food safety and human health, necessitating effective monitoring and assessment techniques. Recognizing characteristic spectral features is essential for accurately assessing the ecological risks associated with heavy metal pollution. To explore the potential of utilizing the hyperspectrum to evaluate ecologial risks of heavy metals pollution in paddy soil, the study investigated ecological risks and spectral characteristics (350 ~ 2500 nm) of heavy metals (Cd, Pb, Cr, Cu) in rice paddy soil in Xiangtan County, Hunan Province, China. Potential ecological risk index (RI) was employed to evaluate pollution level, and derivative transformations (first derivative R′ and second derivative R″) were applied to eliminate the hyperspectral noise. Results identified Cd as the primary pollutant in the paddy soil of Xiangtan County, followed by Pb, Cu, and Cr. Serious heavy metal pollution concentrated in riverine and cropland areas. After applying spectral transformations, the correlation between spectral reflectance and ecological risks significantly improved. The R″ transformation demonstrated the highest correlations for Cd (r = 0.95) at 2046 nm in high ecological risk category, while the R′ transformation exhibited the strongest negative correlation for Cd (r = − 0.94) at 2059 nm. This study highlights the potential of hyperspectral analysis as an effective tool for estimating ecological risks posed by heavy metal pollution in rice paddy soils. The findings provide valuable insights for improving monitoring practices and informing risk management strategies in agricultural environments.
- Open Access
- 28 December 2023
- Published by Mongolian Journals Online in Mongolian Journal of Geography and Geoecology
- Vol. 60 (44) , 222-230
Soil moisture mapping using machine learning technique
Soil moisture is an essential component in the energy cycle, water resource, hydrological regime, and processes of the land surface. Mapping and monitoring of soil moisture are crucial for the prevention of flood and drought, estimation of evapotranspiration, and water resource management. Using remote sensing to create soil moisture mapping at large scale has become one of the most energy and time-efficient methods in soil study. Thus, we aimed to map the soil moisture for Mongolia based on downscaled Soil Moisture Active Passive (SMAP) data by combining it with the Normalized Difference Vegetation Index (NDVI), Enhanced Vegetation Index (EVI), and Land Surface Temperature (LST) of Moderate Resolution Imaging Spectroradiometer (MODIS) data using the Machine Learning-based Random Forest (RF) approach. The SMAP was positively correlated with NDVI (r=0.72, p<0.01) and EVI (r=0.73, p<0.01) but it was negatively correlated with LST (r= -0.66, p<0.05). The performance of the RF was high, and the correlation was r2=0.7. Therefore, our study suggests that the Machine Learning-based RF approach can be used to model soil moisture on a large scale. Машин сургалтын аргаар хөрсний чийгийг зураглах арга зүй Хөрсний чийг нь усны эргэлт, энергийн урсгалд чухал нөлөө үзүүлдгээс гадна, газрын гадаргын нөхцөл болон гадаргын усанд маш чухал нөлөөтэй. Иймд, хөрсний чийгийн зураглал болон мониторингийн судалгаа нь ган, зудын мониторинг, үерийн урьдчилсан сэрэмжлүүлэг болон усны нөөцийн менежментэд чухал үүрэг гүйцэтгэдэг судалгааны нэг юм. Сүүлийн үед, өргөн уудам газар нутагт хөрсний чийгийг зураглахын тулд зайнаас тандан судлалын аргыг ашиглах нь эдийн засаг болон цаг хугацааны хувьд үр ашигтай аргуудын нэг болоод байна. Иймд Монгол орны хэмжээнд хөрсний чийгийг зураглахдаа Soil Moisture Active Passive (SMAP) хиймэл дагуулын бүтээгдэхүүнийг ашиглан машин сургалтын санамсаргүй ой (RF)-н аргаар мэдээний орон зайн шийдийг сайжруулан зураглалаа. Ингэхдээ Moderate Resolution Imaging Spectroradiometer (MODIS) хиймэл дагуулын бүтээгдэхүүнүүдэд (ургамлын нормчилсон ялгаврын индекс (NDVI), ургамлын сайжруулсан индекс (EVI), газрын гадаргын температур (LST) тулгуурлан SMAP хиймэл дагуулын бүтээгдэхүүний орон зайн шийдийг сайжруулан өөрчилсөн, хамаарлыг тооцсон. Ингэхэд NDVI (r=0.72, p<0.01) болон EVI (r=0.73, p<0.01) нь SMAP-тай эерэг хамааралтай байсан бол LST (r= -0.66, p<0.05)-тай урвуу хамааралтай байсан. RF-н алгоритмаар машин сургалтын аргыг ашиглан Монгол орны хэмжээнд хөрсний чийгийг зураглахад загварын үр дүн гүйцэтгэл сайтай буюу хамаарал нь r2=0.7 гарсан. Иймд машин сургалтын санамсаргүй ойн алгоритмаар том хэмжээний газар нутгийг хамруулан хөрсний чийгийг загварчлах боломжтой болох нь судалгааны үр дүнгээс харагдаж байна. Түлхүүр үгс: Хөрсний чийг, машин сургалт, SMAP
- 28 December 2023
- Published by Mongolian Journals Online in Mongolian Journal of Geography and Geoecology
- Vol. 60 (44) , 222-230
Open AccessSoil moisture mapping using machine learning technique
Soil moisture is an essential component in the energy cycle, water resource, hydrological regime, and processes of the land surface. Mapping and monitoring of soil moisture are crucial for the prevention of flood and drought, estimation of evapotranspiration, and water resource management. Using remote sensing to create soil moisture mapping at large scale has become one of the most energy and time-efficient methods in soil study. Thus, we aimed to map the soil moisture for Mongolia based on downscaled Soil Moisture Active Passive (SMAP) data by combining it with the Normalized Difference Vegetation Index (NDVI), Enhanced Vegetation Index (EVI), and Land Surface Temperature (LST) of Moderate Resolution Imaging Spectroradiometer (MODIS) data using the Machine Learning-based Random Forest (RF) approach. The SMAP was positively correlated with NDVI (r=0.72, p<0.01) and EVI (r=0.73, p<0.01) but it was negatively correlated with LST (r= -0.66, p<0.05). The performance of the RF was high, and the correlation was r2=0.7. Therefore, our study suggests that the Machine Learning-based RF approach can be used to model soil moisture on a large scale. Машин сургалтын аргаар хөрсний чийгийг зураглах арга зүй Хөрсний чийг нь усны эргэлт, энергийн урсгалд чухал нөлөө үзүүлдгээс гадна, газрын гадаргын нөхцөл болон гадаргын усанд маш чухал нөлөөтэй. Иймд, хөрсний чийгийн зураглал болон мониторингийн судалгаа нь ган, зудын мониторинг, үерийн урьдчилсан сэрэмжлүүлэг болон усны нөөцийн менежментэд чухал үүрэг гүйцэтгэдэг судалгааны нэг юм. Сүүлийн үед, өргөн уудам газар нутагт хөрсний чийгийг зураглахын тулд зайнаас тандан судлалын аргыг ашиглах нь эдийн засаг болон цаг хугацааны хувьд үр ашигтай аргуудын нэг болоод байна. Иймд Монгол орны хэмжээнд хөрсний чийгийг зураглахдаа Soil Moisture Active Passive (SMAP) хиймэл дагуулын бүтээгдэхүүнийг ашиглан машин сургалтын санамсаргүй ой (RF)-н аргаар мэдээний орон зайн шийдийг сайжруулан зураглалаа. Ингэхдээ Moderate Resolution Imaging Spectroradiometer (MODIS) хиймэл дагуулын бүтээгдэхүүнүүдэд (ургамлын нормчилсон ялгаврын индекс (NDVI), ургамлын сайжруулсан индекс (EVI), газрын гадаргын температур (LST) тулгуурлан SMAP хиймэл дагуулын бүтээгдэхүүний орон зайн шийдийг сайжруулан өөрчилсөн, хамаарлыг тооцсон. Ингэхэд NDVI (r=0.72, p<0.01) болон EVI (r=0.73, p<0.01) нь SMAP-тай эерэг хамааралтай байсан бол LST (r= -0.66, p<0.05)-тай урвуу хамааралтай байсан. RF-н алгоритмаар машин сургалтын аргыг ашиглан Монгол орны хэмжээнд хөрсний чийгийг зураглахад загварын үр дүн гүйцэтгэл сайтай буюу хамаарал нь r2=0.7 гарсан. Иймд машин сургалтын санамсаргүй ойн алгоритмаар том хэмжээний газар нутгийг хамруулан хөрсний чийгийг загварчлах боломжтой болох нь судалгааны үр дүнгээс харагдаж байна. Түлхүүр үгс: Хөрсний чийг, машин сургалт, SMAP