Predicting landslide susceptibility based on decision tree machine learning models under climate and land use changes DOI
Quoc Bao Pham, Subodh Chandra Pal, Rabin Chakrabortty

et al.

Geocarto International, Journal Year: 2021, Volume and Issue: 37(25), P. 7881 - 7907

Published: Sept. 27, 2021

Landslides are most catastrophic and frequently occurred across the world. In mountainous areas of globe, recurrent occurrences landslide have caused huge amount economic losses a large number casualties. this research, we attempted to estimate potential impact climate LULC on future susceptibility in Markazi Province Iran. We considered boosted tree (BT), random forest (RF) extremely randomized (ERT) models for assessment Province. The results evaluation criteria showed that ERT model is optimal than other used study with AUC values 0.99 0.93 training validation datasets, respectively. According model, spatial coverage very high land slide susceptible zones current period, 2050s considering RCP 2.6 8.5 428.5 km2, 439.6 km2 465.2 From analysis it clear changes prominent. present help managers reduce damages, not only but also conditions, based changes.

Language: Английский

GIS and remote sensing-based assessment of soil erosion risk using RUSLE model in South-Kivu province, eastern, Democratic Republic of Congo DOI Creative Commons
Luc Cimusa Kulimushi, Pandurang Choudhari, Léonard Mubalama

et al.

Geomatics Natural Hazards and Risk, Journal Year: 2021, Volume and Issue: 12(1), P. 961 - 987

Published: Jan. 1, 2021

Soil erosion risk assessment in South-Kivu longs for the colonial epoch, while province faces problem of extreme degradation land form soil erosion. Thus, study attempts to assess at level using Revised Universal Loss Equation (RUSLE) conjunction with Geographical Information System (GIS), and remote sensing data. The estimated total was 2.084 million tons; an annual average 138.2 t ha−1 yr−1. Moreover, loss greater than 100 yr−1 accounts 45.2% erosive land. worsening nearly entire territories range between 87 Shabunda 248 Uvira. Under high aggressiveness rainfall mean 1857.19 mm/y, highest rate found Perennial crop, Trees, Cropland contrast Shrub closed Forest mainly due slope 22% former Land cover categories compared 17% Shrubland forest. adoption terracing could reduce by 76% current cropland i.e., from (162.12 38 yr−1). Therefore it is strongly recommended.

Language: Английский

Citations

53
Towards improved USLE-based soil erosion modelling in India: A review of prevalent pitfalls and implementation of exemplar methods DOI
Anindya Majhi, Rohit Shaw, Kunal Mallick

et al.

Earth-Science Reviews, Journal Year: 2021, Volume and Issue: 221, P. 103786 - 103786

Published: Aug. 28, 2021

Language: Английский

Citations

52
Spatial prediction of landslide susceptibility using projected storm rainfall and land use in Himalayan region DOI
Indrajit Chowdhuri, Subodh Chandra Pal, Rabin Chakrabortty

et al.

Bulletin of Engineering Geology and the Environment, Journal Year: 2021, Volume and Issue: 80(7), P. 5237 - 5258

Published: May 2, 2021

Language: Английский

Citations

50
Deep learning and boosting framework for piping erosion susceptibility modeling: spatial evaluation of agricultural areas in the semi-arid region DOI
Yunzhi Chen, Wei Chen, Saeid Janizadeh

et al.

Geocarto International, Journal Year: 2021, Volume and Issue: 37(16), P. 4628 - 4654

Published: Feb. 19, 2021

Piping erosion is one of the water erosions that cause significant changes in landscape, leading to environmental degradation. To prevent losses resulting from tube growth and enable sustainable development, developing high-precision predictive algorithms for piping essential. Boosting a classic algorithm has been successfully applied diverse computer vision tasks. Therefore, this work investigated performance Boosted Linear Model (BLM), Regression Tree (BRT), Generalized (Boost GLM), Deep models susceptibility mapping Zarandieh Watershed located Markazi province Iran. A inventory map including 152 locations was prepared training testing. 18 initial predisposing factors (altitude, slope, plan curvature, profile distance river, drainage density, road, rainfall, land use, soil type, bulk CEC, pH, clay, silt, sand, topographical position index (TPI), topographic wetness (TWI)) derived multiple remote sensing (RS) sources determine prone areas. The most were selected using multi-collinearity analysis which indicates linear correlations between factors. Finally, results evaluated Sensitivity, Specificity, Positive values (PPV) Negative value (NPV), Receiver Operation characteristic (ROC) curve. best Sensitivity (0.80), Specificity (0.84), PPV (0.85), NPV (0.79), ROC (0.93), obtained by model. study agricultural use showed 41% lands are very sensitive erosion. This outcome will natural resource managers local planners assess take effective decisions minimize damages accurately identifying vulnerable Hence, research proved model's ability comparison other popular methods such as BLM, BRT, Boost GLM.

Language: Английский

Citations

48
Predicting landslide susceptibility based on decision tree machine learning models under climate and land use changes DOI
Quoc Bao Pham, Subodh Chandra Pal, Rabin Chakrabortty

et al.

Geocarto International, Journal Year: 2021, Volume and Issue: 37(25), P. 7881 - 7907

Published: Sept. 27, 2021

Landslides are most catastrophic and frequently occurred across the world. In mountainous areas of globe, recurrent occurrences landslide have caused huge amount economic losses a large number casualties. this research, we attempted to estimate potential impact climate LULC on future susceptibility in Markazi Province Iran. We considered boosted tree (BT), random forest (RF) extremely randomized (ERT) models for assessment Province. The results evaluation criteria showed that ERT model is optimal than other used study with AUC values 0.99 0.93 training validation datasets, respectively. According model, spatial coverage very high land slide susceptible zones current period, 2050s considering RCP 2.6 8.5 428.5 km2, 439.6 km2 465.2 From analysis it clear changes prominent. present help managers reduce damages, not only but also conditions, based changes.

Language: Английский

Citations

48