edoc

Soil erosion risk map for Swiss grasslands : a dynamic approach to model the spatio-temporal patterns of soil loss

Schmidt, Simon. Soil erosion risk map for Swiss grasslands : a dynamic approach to model the spatio-temporal patterns of soil loss. 2018, Doctoral Thesis, University of Basel, Faculty of Science.

[img]
Preview
PDF
Available under License CC BY-NC (Attribution-NonCommercial).

19Mb

Official URL: http://edoc.unibas.ch/diss/DissB_13239

Downloads: Statistics Overview

Abstract

Soil erosion by water on grassland does not attract the same attention like erosion on arable land as it is usually assumed that the closed vegetation cover prevents soil loss. However, the complex terrain and intensive pasture use of mountain grasslands can potentially induce high soil loss. With a share of 72% of the total agricultural area, grassland is one of the most dominant land use in Switzerland and therefore should not be neglected in topics concerning soil protection.
Previous soil erosion studies revealed that soil erosion rates in Switzerland are not constant over time but rather are highly dynamic within a year. Such seasonal variability is mainly caused by rainfall patterns and plant growth cycles. Hence, modeling of soil loss based on a seasonal resolution enables improved insights in the erosion dynamics within a year.
The present work aims to model soil erosion with a sub-annual resolution for Swiss grasslands. Thereby we will focus on the most dynamic soil erosion risk factors namely rainfall erosivity and land cover and management.
The soil erosion model itself relies on the Revised Universal Soil Loss Equation (RUSLE). Each of the erosion factors of the RUSLE (rainfall erosivity R, soil erodibility K, cover and management C, slope length L, slope steepness S, and support practices P) is modified according to the specific environmental conditions of Swiss grasslands. The factors R and C are the most variable factors within a year as they are directly related to the parameters rainfall intensity and plant growth cycle. Therefore, both factors are modeled on a monthly scale to capture the temporal variations of soil loss within the year. For flexibility and transparency reasons, we derived each factor separately with the most state-of-the-art data and methodology as each of the factor transmit information about its effect on the overall model. Support practices (P-factor) are not considered in the model as the parametrization of grassland management practices and their effect for erosion control is difficult due to a lack of data and studies.
Monthly estimates of the rainfall erosivity (R-factor) are based on 10-minutes rainfall data of 87 gauging stations distributed all over Switzerland. Subsequently, the monthly rainfall erosivity is interpolated with spatial covariates representing snow cover, precipitation, and topography. For the C-factor, the fraction of green vegetation cover (FGVC) was derived from the 0.25 m spatial resolution Swissimage orthophotos by a linear spectral unmixing technique. A temporal normalization of the spatial distribution of the FGVC combined with R-factor weighting results in spatial and temporal patterns of the C-factor. Soil erodibility (expressed as the K-factor of the RUSLE equation) was modeled with cubist regression and multilevel B-splines on a national scale based on a total of 199 Swiss and 1639 European Land Use/Cover Area frame statistical Survey (LUCAS) topsoil samples. The LS-factor was adopted to the steep alpine environment by limiting the slope length to 100 m and using a fitted S-factor of empirical slope steepness factors.
The mean monthly modeled R-factor for Switzerland is 96.5 MJ mm ha-1 h-1 month-1. On average, rainfall erosivity is 25 times higher in August (263.5 MJ mm ha-1 h-1 month-1) then in January (10.5 MJ mm ha-1 h-1 month-1). In general, the winter has relatively low R-factor values (average of 14.7 MJ mm ha-1 h-1 month-1). The mean monthly C-factor on Swiss grasslands is 0.012 with a maximum from May until September. The national average K-factor of Switzerland is 0.0327 t ha h ha-1 MJ-1 mm-1. The LS-factor for Switzerland is relatively high (14.8) compared to other countries but is mainly driven by the complex topography of the Alps with its steep slopes.
The soil erosion modeling reveals distinct seasonal variations. July and August are identified to be the months with the highest soil loss rates (1.25 t ha-1 month-1) by water on Swiss grasslands. Spatially, hotspots of soil erosion are in the Central Swiss Alps (parts of the cantons Fribourg, Bern, Obwalden, Nidwalden, St. Gallen, Appenzell Innerrhoden, and Appenzell Ausserrhoden) in summer. Winter is the season with the lowest risk of soil loss due to low rainfall erosivity on snow-covered ground. The average annual soil loss for Switzerland, expressed as the sum of all monthly erosion rates, is 4.55 t ha-1 yr-1.
The spatial rainfall erosivity patterns are heterogeneous in all months, but spatial differences are less pronounced in winter due to the low rainfall erosivity. The small-scale variability of rainfall erosivity is less distinct in all months as homogenous rainfall patterns usually cover larger regions controlled mainly by topography. However, the Swiss Alps are not equally affected by rainfall erosivity with a very low variability within a year in the western and eastern Alps. In contrast, the small-scale variability of the cover and management factor is higher in most of the months due to the impact of grassland land use. The average C-factor for Swiss grassland of 0.012 matches the commonly applied C-factor for grasslands (0.01) proposed in the literature. The Swiss K-factor is low to medium with a clear reduction under consideration of the surface stone cover. We expected a high LS-factor for Switzerland as steep slopes are frequently in the Swiss Alps.
The dominance of soil erosion risk on grasslands in summer is surprising as it is commonly assumed that the closed vegetation cover protects soils. Though, the individual consideration of all factors, especially of the R- and C-factor, reveal their strong effect and interaction within the erosion model. The average annual soil loss prediction for Swiss grassland exceeds the maximum tolerable soil loss of Switzerland (2 t ha-1 yr-1; Schaub and Prasuhn, 1998) by a factor of 2. That modeling result highlights that soil erosion on grasslands is of high concern for the Swiss agricultural productivity and environmental protection of a large proportion of the Swiss territory.
Based on the increased temporal resolution of soil erosion predictions, spatial and temporal patterns of soil loss by water on Swiss grasslands can be captured. The simultaneous identification of spatial and temporal patterns of soil loss on Swiss grasslands makes a targeted soil erosion control feasible. The knowledge about where and when soil erosion occurs enables the implementation of selective erosion control measures specifically for time periods and regions with high susceptibility.
Developing a comprehensive soil erosion assessment on Swiss grassland that is comparable and connectable with available risk assessments such as the erosion risk map 2 for Swiss arable lands (Prasuhn et al., 2013) and the European Union’s assessment RUSLE2015 (Panagos et al., 2015) provides a national and even continental valuation of soil erosion risk. The soil erosion risk map can be seen as a prototype for other erosion modeling on grassland in the Alpine region.
Advisors:Alewell, Christine and Klaus, Julian
Faculties and Departments:05 Faculty of Science > Departement Umweltwissenschaften > Geowissenschaften > Umweltgeowissenschaften (Alewell)
UniBasel Contributors:Schmidt, Simon and Alewell, Christine
Item Type:Thesis
Thesis Subtype:Doctoral Thesis
Thesis no:13239
Thesis status:Complete
Number of Pages:1 Online-Ressource (XII, 168 Seiten, Supplement 169-Supplement 195)
Language:English
Identification Number:
edoc DOI:
Last Modified:23 Jul 2019 04:30
Deposited On:22 Jul 2019 13:57

Repository Staff Only: item control page