Modelling actual and potential wind erosion risk by using readily available data on weather elements and GIS : "a pilot study from Denmark and Switzerland"

Saremi Naeini, Mohammadali. Modelling actual and potential wind erosion risk by using readily available data on weather elements and GIS : "a pilot study from Denmark and Switzerland". 2015, Doctoral Thesis, University of Basel, Faculty of Science.

Available under License CC BY-NC-ND (Attribution-NonCommercial-NoDerivatives).


Official URL: http://edoc.unibas.ch/diss/DissB_11435

Downloads: Statistics Overview


Wind erosion is a complex process which is affected generally by the combined impact of wind erosivity and soil erodibility. According to the complexity of the wind erosion process, the main aim of this research was to design a practical model to predict potential and actual wind erosion risk based on spatial distribution of wind erosivity and soil erodibility. For representation of the potential and actual wind erosion risk in this study, two pilot countries with very different environmental settings were selected; Denmark and Switzerland. In order to be able to implement the model even in areas with minimal data availability, the structure of the GIS-based model was designed for a limited number of key parameters, which can be easily accessed even in regions with inadequate data. Three basic aspects distinguish the proposed model from other models:
1) Separation of wet- and dry-times is taken into account for the wind data analysis;
2) The impact of climate change is considered for factors that are used in the model;
3) Running the model for given return periods based on extreme wind velocity analysis.
The soil moisture content is one of the most important and dynamic factors determining soil resistance to wind erosion, because it affects threshold wind velocities for particle detachment. Presence or absence of moisture in the soil should therefore, be included in wind erosion risk assessments. In order to include soil moisture conditions into the wind data analysis, a sub-model was developed to separate wet and dry periods in weather time series. Weather data and soil moisture content collected during one year in Foulum were used to calibrate the model.
The reason why soil moisture conditions were considered for the wind data analysis was the theory that using wind data for calculation of wind frequency distributions regardless of the status of soil moisture would lead to an overestimation of wind erosion. To confirm or reject this hypothesis, the frequency distribution of wind velocity in conventional method (all-times) and proposed method (dry-times) was compared by using the pairwise Wilcoxon test. The results showed that in 99.6% and 97.1% of tests the difference between two distributions were significant at 99% confidence level for Denmark and Switzerland respectively. Change detection analysis of wind erosivity maps, revealed that 56.02% and 30.63% of the territory of Denmark and Switzerland experience an overestimation of wind erosivity, if the conventional approach would be used. However, underestimation was also observes in considerable parts of these countries, but almost all of these areas were located in regions, which are not prone to wind erosion.
To investigate the impact of climate change on various wind factors, the Mann-Kendall trend test and the Sen’s slope estimator method for detecting the trend and estimating its magnitude were used. The results revealed that, in general, most wind factors experience a slightly decreasing trend in both countries. The median of trends of each input factor was considered to assess the impact of climate change on wind erosion risk modeling.
For running the model according to a given return period it was necessary to analyzed extreme wind velocities and to extract return levels for desired return periods. For this goal the Peak Over Threshold (POT) method was considered and the time series was fitted by the Generalized Pareto distribution (GPD) model. To ensure the independence of the extracted extreme values, a peak over threshold identifier algorithm was designed based on the detection of windiness of periods.
According to the results of above mentioned investigations, a GIS-based model was designed and successfully implemented to generate spatio-temporal distributions of potential and actual wind erosion risk by using a two-dimensional minimum curvature spline technique to spatially interpolate data, as well as a fuzzy overlay technique to combine fuzzy membership rasters.
The results of model in Switzerland confirmed that, wind erosion is not a threat in this country and only in limited areas of croplands (1.7% of croplands) the risk of wind erosion was estimated to be high. In Denmark 11.48% of total land surface was ranked in the class of high actual wind erosion risk, which are generally located in the north-west and south-west of Jutland peninsula as well as north of Vendsyssel-Thy and Zealand.
The spatial distribution of actual wind erosion risk in Denmark revealed that almost all of the estimated high risks occurs in “Croplands” and “barren or sparsely vegetated” lands, which included 18.4% and 10.3% of these land types respectively. Therefore, it should be emphasized that the role of human activities can have a significant impact on the increase or decrease of wind erosion risk in both countries.
Advisors:Fister, Wolfgang
Committee Members:Casper, Markus C.
Faculties and Departments:05 Faculty of Science > Departement Umweltwissenschaften > Geowissenschaften
UniBasel Contributors:Fister, Wolfgang
Item Type:Thesis
Thesis Subtype:Doctoral Thesis
Thesis no:11435
Thesis status:Complete
Number of Pages:208 S.
Identification Number:
edoc DOI:
Last Modified:08 Feb 2020 14:08
Deposited On:15 Oct 2015 10:22

Repository Staff Only: item control page