Ecological suitability assessment methods of waste pile-up along railway routes based on machine learning algorithms DOI Creative Commons
Cuicui Ji,

Zaoyang Huang,

Xiangjun Pei

et al.

Ecosystem Health and Sustainability, Journal Year: 2024, Volume and Issue: 10

Published: Jan. 1, 2024

Waste pile-up along railway routes poses an important threat to the regional ecological environment. However, there is a lack of methods that assess suitability waste (ESWP) at macro scale, which crucial for informed decision-making. We define ESWP and propose methodology measure level routes. Specifically, we focus on Ya’an Nyingchi section railway, selecting 30-km buffer zone either side as study area. To develop maps, employed Landsat 8, digital elevation model (DEM), soil database, land use, meteorological data. tested 3 machine learning methods—random forest (RF), deep neural network (DNN), extreme gradient boosting (XGBoost)—using 7 key indicators input parameters. The performance these models was evaluated using overall accuracy Kappa index. Additionally, analyzed relative importance each indicator results. reached following results: Firstly, combination selected with effectively assesses railways. Secondly, among methods, DNN demonstrated superior performance, achieving 86.49%, outperforming RF (80.31%) XGBoost (79.54%). Thirdly, greatest impact assessment were biological richness (weight 0.23), vegetation coverage 0.20), nutrients 0.16). These findings provide novel approach assessing identifying low-risk sites

Language: Английский

Soil and Water Assessment Tool-Based Prediction of Runoff Under Scenarios of Land Use/Land Cover and Climate Change Across Indian Agro-Climatic Zones: Implications for Sustainable Development Goals DOI Open Access
Subbarayan Saravanan, Youssef M. Youssef, Leelambar Singh

et al.

Water, Journal Year: 2025, Volume and Issue: 17(3), P. 458 - 458

Published: Feb. 6, 2025

Assessing runoff under changing land use/land cover (LULC) and climatic conditions is crucial for achieving effective sustainable water resource management on a global scale. In this study, the focus was predictions across three diverse Indian watersheds—Wunna, Bharathapuzha, Mahanadi—spanning distinct agro-climatic zones to capture varying hydrological complexities. The soil assessment (SWAT) tool used simulate future influenced by LULC climate change explore related sustainability implications, including challenges proposing countermeasures through action plan (SAP). methodology integrated high-resolution satellite imagery, cellular automata (CA)–Markov model projecting changes, downscaled data representative concentration pathways (RCPs) 4.5 8.5, representing moderate extreme scenarios, respectively. SWAT calibration validation demonstrated reliable predictive accuracy, with coefficient of determination values (R2) > 0.50 confirming reliability in simulating processes. results indicated significant increases surface due urbanization, reaching >1000 mm, 600 400 mm southern southeastern Wunna, northwestern Mahanadi, respectively, especially 2040 RCP 8.5. These findings indicate that quality, agricultural productivity, urban infrastructure may be threatened. proposed SAP includes nature-based solutions, like wetland restoration, climate-resilient strategies mitigate adverse effects partially achieve development goals (SDGs) clean action. This research provides robust framework watershed similar regions worldwide.

Language: Английский

Citations

5

Uğursuyu Havzası Erozyon Risk Durumundaki Dönemsel Değişimlerin Belirlenmesi DOI Open Access
Ahmet Salih Değermenci

Bartın Orman Fakültesi Dergisi, Journal Year: 2025, Volume and Issue: 27(1), P. 15 - 32

Published: April 21, 2025

Bu çalışmada, Uğursuyu Havzasında 2000 ve 2019 yılları arasında arazi kullanım durumlarında meydana gelen değişikliklerin erozyon risk durumlarına etkileri ICONA (National Institute for Nature Conservation) modeli kullanılarak detaylı bir şekilde incelenmiştir. Arazi sınıflandırmasında, su, yerleşim, tarım-açıklık bitki örtüsü olmak üzere dört ana sınıfı belirlenmiş bu sınıfların doğruluğu hata matrisi yönt-emiyle değerlendirilmiştir. Kappa değerleri, her iki dönem için %80’in üzerinde bulunmuş, da sınıflandırmanın oldukça başarılı olduğunu göstermiştir. Su alanları 14,86 ha’dan 18,05 ha’a yükselirken, yerleşim alanlarında yaklaşık 100 ha’lık artış gözlemlenmiştir. Bununla birlikte, alanlarının oranı %84,6’dan %72,3’e düşmüştür. Toprak koruma haritaları, ile oranları hari-talarının ilişkilendirilmesiyle oluşturulmuş süreçte çok düşük yüksek toprak sınıflarında artışlar gözlemlenirken, orta azalmalar gelmiştir. Havzanın eğim jeolojik yapısı dikkate alınarak hazırlanan potansiyel alanın %76,5’inin riski grubunda bulunduğunu göstermektedir. Jeolojik yapı olarak büyük kısmı (%80,5) erozyona duyarlı kayaçlardan oluşmaktadır. Erozyon durumları açısından yapılan analizlerde, önemli değişim gözlemlenmezken, seviyede %1’lik %3,6’lık azalma Çok sınıfında ise %2,54'lük kaydedilmiştir. değişimleri oranlarındaki azalmalar, riskini etkileyen temel faktörler öne çıkmıştır. modeli, etkili değerlendirmiş havzanın duyarlılığına sahip ortaya koymuştur. Sonuç olarak, elde edilen bulgular, sürdürülebilir yönetimi kontrolünün önemini vurgulamaktadır. bağlamda, yerel yönetimlerin toplulukların iş birliği çevresel önlemlerinin alınması, bölgenin ekolojik dengesinin korunması kritik ger-eklilik haline gelmektedir.

Citations

0

Ecological suitability assessment methods of waste pile-up along railway routes based on machine learning algorithms DOI Creative Commons
Cuicui Ji,

Zaoyang Huang,

Xiangjun Pei

et al.

Ecosystem Health and Sustainability, Journal Year: 2024, Volume and Issue: 10

Published: Jan. 1, 2024

Waste pile-up along railway routes poses an important threat to the regional ecological environment. However, there is a lack of methods that assess suitability waste (ESWP) at macro scale, which crucial for informed decision-making. We define ESWP and propose methodology measure level routes. Specifically, we focus on Ya’an Nyingchi section railway, selecting 30-km buffer zone either side as study area. To develop maps, employed Landsat 8, digital elevation model (DEM), soil database, land use, meteorological data. tested 3 machine learning methods—random forest (RF), deep neural network (DNN), extreme gradient boosting (XGBoost)—using 7 key indicators input parameters. The performance these models was evaluated using overall accuracy Kappa index. Additionally, analyzed relative importance each indicator results. reached following results: Firstly, combination selected with effectively assesses railways. Secondly, among methods, DNN demonstrated superior performance, achieving 86.49%, outperforming RF (80.31%) XGBoost (79.54%). Thirdly, greatest impact assessment were biological richness (weight 0.23), vegetation coverage 0.20), nutrients 0.16). These findings provide novel approach assessing identifying low-risk sites

Language: Английский

Citations

0