Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2020 Nov 24;10(1):20494.
doi: 10.1038/s41598-020-77567-0.

Mapping wind erosion hazard with regression-based machine learning algorithms

Hamid Gholami  1 , Aliakbar Mohammadifar  2 , Dieu Tien Bui  3   4 , Adrian L Collins  5
Affiliations

Mapping wind erosion hazard with regression-based machine learning algorithms

Hamid Gholami et al. Sci Rep. .

Abstract

Land susceptibility to wind erosion hazard in Isfahan province, Iran, was mapped by testing 16 advanced regression-based machine learning methods: Robust linear regression (RLR), Cforest, Non-convex penalized quantile regression (NCPQR), Neural network with feature extraction (NNFE), Monotone multi-layer perception neural network (MMLPNN), Ridge regression (RR), Boosting generalized linear model (BGLM), Negative binomial generalized linear model (NBGLM), Boosting generalized additive model (BGAM), Spline generalized additive model (SGAM), Spike and slab regression (SSR), Stochastic gradient boosting (SGB), support vector machine (SVM), Relevance vector machine (RVM) and the Cubist and Adaptive network-based fuzzy inference system (ANFIS). Thirteen factors controlling wind erosion were mapped, and multicollinearity among these factors was quantified using the tolerance coefficient (TC) and variance inflation factor (VIF). Model performance was assessed by RMSE, MAE, MBE, and a Taylor diagram using both training and validation datasets. The result showed that five models (MMLPNN, SGAM, Cforest, BGAM and SGB) are capable of delivering a high prediction accuracy for land susceptibility to wind erosion hazard. DEM, precipitation, and vegetation (NDVI) are the most critical factors controlling wind erosion in the study area. Overall, regression-based machine learning models are efficient techniques for mapping land susceptibility to wind erosion hazards.

PubMed Disclaimer

Conflict of interest statement

The authors declare no competing interests.

Figures

Figure 1
Figure 1
The relative importance of the effective factors for wind erosion estimated by MMLPNN. DEM, PR, NDVI, AWC, CCP, ESP, OCC, EC, GE, BD, WS, ST, and LU indicate digital elevation model, precipitation, normalized difference vegetation index, available water content, calcium carbonate content, exchangeable sodium percentage, organic carbon content, electrical conductivity, geology, bulk density, wind speed, soil texture, and land use, respectively.
Figure 2
Figure 2
Maps of wind erosion hazard generated by: (a) RLR, (b) Cforest, (c) NCPQR, (d) NNFE, (e) MMLPNN, and (f) RR. The values for pixels was estimated by R software (https://CRAN.R-project.org/doc/FAQ/R-FAQ.html) and then, values of pixels were mapped by ArcGIS 10.4.1 (https://www.esri.com/en-us/about/about-esri/overview).
Figure 3
Figure 3
Maps of wind erosion hazard generated by: (a) BGLM, (b) NBGLM, (c) BGAM, (d) SGAM, (e) SSR, and (f) SGB. The values for pixels was estimated by R software (https://CRAN.R-project.org/doc/FAQ/R-FAQ.html) and then, values of pixels were mapped by ArcGIS 10.4.1 (https://www.esri.com/en-us/about/about-esri/overview).
Figure 4
Figure 4
Maps of wind erosion hazard generated by: (a) SVM, (b) RVM, (c) Cubist and (d) ANFIS. The values for pixels was estimated by R software (https://CRAN.R-project.org/doc/FAQ/R-FAQ.html) and then, values of pixels were mapped by ArcGIS 10.4.1 (https://www.esri.com/en-us/about/about-esri/overview).
Figure 5
Figure 5
The values of the statistical indicators were used to evaluate model performance; (a) training dataset and (b) evaluation dataset.
Figure 6
Figure 6
Taylor diagrams for assessing the performance of the models in this research; (a) training dataset, and (b) evaluation dataset.
Figure 7
Figure 7
Location of the study area in Iran and sampling sites used for this study. Soil sampling sites were extracted from the world soil map and then, these sites were mapped in ArcGIS 10.4.1 (https://www.esri.com/en-us/about/about-esri/overview).
Figure 8
Figure 8
Spatial maps of soil characteristics: (a) AWC; (b) bulk density; (c) calcium carbonate percentage; (d) EC; (e) ESP, and; (f) OCC. All these factors were mapped spatially in ArcGIS 10.4.1 (https://www.esri.com/en-us/about/about-esri/overview).
Figure 9
Figure 9
Spatial maps of: (a) soil texture; (b) geology; (c) land use; (d) NDVI; (e) DEM, and; (f) wind speed. All these factors were mapped spatially in ArcGIS 10.4.1 (https://www.esri.com/en-us/about/about-esri/overview).
Figure 10
Figure 10
Spatial maps of: (a) total annual precipitation; (b) locations of the pixels with active wind erosion, and; (c) locations of the training and validation data points. All these characteristics were mapped spatially in ArcGIS 10.4.1 (https://www.esri.com/en-us/about/about-esri/overview).
Figure 11
Figure 11
Flowchart of the methodology for mapping of wind erosion hazard.
See this image and copyright information in PMC

Similar articles

See all similar articles

Cited by

References

    1. Prospero JM, Ginoux P, Torres O, Nicholson SE, Gill TE. Environmental characterization of global sources of atmospheric soil dust identified with the Nimbus 7 Total Ozone Mapping Spectrometer (TOMS) absorbing aerosol product. Rev. Geophys. 2002;40(1):1–31. doi: 10.1029/2000RG000095. - DOI
    1. Goossens D. On-site and off-site effects of wind erosion. In: Warren A, editor. Wind Erosion on Agricultural Land in Europe. European Commission: Luxembourg; 2003. pp. 29–38.
    1. Dahmardeh Behrooz, R., Gholami, H., Telfer, M. W., Jansen, J. D. & Fathabadi, A. Uisng GLUE to pull apart the provenance of atmospheric dust. Aeolian Res.37, 1–13 (2019).
    1. Collins, A. L., Blackwell, M., Boeckx, P., Chivers, C. A., Emelko, M., Evrard, O., & Harris, P. Sediment source fingerprinting: benchmarking recent outputs, remaining challenges and emerging themes. J. Soils Sedim. 1–34 (2020). - PMC - PubMed
    1. Rashki A, Kaskaoutis DG, Goudie AS, Kahn RA. Dryness of ephemeral lakes and consequences for dust activity: The case of the Hamoun drainage basin, southeastern Iran. Sci. Total Environ. 2013;463–464:552–564. doi: 10.1016/j.scitotenv.2013.06.045. - DOI - PubMed

Publication types