Long-term trends in rainfall erosivity–analysis of high resolution precipitation time series (1937–2007) from Western Germany

https://doi.org/10.1016/j.agrformet.2012.11.011Get rights and content

Abstract

It is generally assumed that global warming will lead to a more dynamic atmosphere which potentially leads to more frequent high intensity rainfall events in many regions of the world. In consequence, an increase in local flash floods and soil erosion intensity would be expected. This study used one of the very rare long-term (1937–2007) high resolution (≤5-min) data sets of ten stations in Central Europe to analyze long-term trends in summer rainfall erosivity. Furthermore potential changes in frequency and/or magnitude of individual erosive rainfall events, and shifts in seasonality of rainfall erosivity were investigated. The data were intensively tested for consistency and homogeneity and trends were analyzed using linear regressions as well as Mann–Kendall tests. For the period of 1937–2007, a slight, significant increase in summer erosivity (April–November) of 4.4% per decade was observed. This linear trend is much steeper since the early seventies of the last century (1973–2007: an increase of 21.0% per decade). For both periods, the linear trend was confirmed by positive and significant results of the Mann–Kendall test. The increasing trends in summer erosivity resulted from an increasing frequency of erosive events and an increase in magnitude, especially of the largest events. The proof of changes in seasonality is, for methodological reasons, less clear than the overall change in summer erosivity. However, there is a tendency that the period of erosive events was prolonged during the last decades of the observations with comparably higher erosivity between May and July and in October. Depending on adaption strategies of farmers, this changing temporal pattern in erosivity might lead to more pounced erosion events under row crops in spring, and after the harvest of small grains in late summer and autumn. In general, this shift in seasonality seemed to be more important for an increase in erosion potential than an overall slight increase in annual erosivity.

Highlights

► Analysis of 71-year data-set of 5-min precipitation from 10 stations in Central Europe. ► Slight increase in summer erosivity (1937–2007) with steeper trends within the last 35 years. ► Increase in frequency and magnitude of erosive events. ► Erosivity increase in spring and autumn especially amplify potential erosion.

Introduction

It is expected that increasing global warming will have multiple effects on the hydrological cycle (Sivakumar, 2011). In most regions an increase and in some a decrease of precipitation is projected (IPCC, 2007). However, apart from regional differences in rainfall depths, it is generally assumed that global warming will lead to a more dynamic atmosphere which potentially leads to more frequent high intensity rainfall events (Groisman et al., 2004, Nearing et al., 2004). Such potential increase in high intensity rainfall events may lead to a number of (unwanted) side effects, e.g. an increase in local flash floods or muddy floods (Boardman et al., 2003, Verstraeten and Poesen, 2000), and an increase in on-site (Lal, 2001) as well as off-site erosion damages (Bilotta et al., 2007, Haygarth et al., 2006).

To evaluate potential effects of changes in rainfall on erosion, traditionally the annual rainfall erosivity (or R factor) of the Universal Soil Loss Equation (USLE; Wischmeier and Smith, 1960), which combines rainfall intensity and depth, is used. The rainfall erosivity can be understood as a variable representing rainfall energy (derived from rainfall intensity) and surface runoff potential (derived from event-based rainfall depth). Since its introduction in the 1970s in the US, the rainfall erosivity was empirically adapted to other regions (e.g. Larionov, 1993, Schwertmann et al., 1990). Moreover, annual rainfall erosivity calculated as the sum of event-based erosivity, is a valuable proxy variable to evaluate the combined change of high rainfall intensities and depths. Hence, its change is also a generally valuable variable for other hydrological purposes.

Climate data derived from a combination of Global Circulation Models and different statistical or dynamical downscaling methods hardly provide the necessary detailed storm information (time step  10 min) needed to calculate rainfall erosivity. Therefore, existing evaluations of changes in rainfall erosivity are relatively rare and are based on (i) some long-term, high resolution rainfall data (e.g. Meusburger et al., 2012, Verstraeten et al., 2006) and/or (ii) empirical relations between rainfall erosivity and monthly to yearly rainfall depth (e.g. Diodato and Bellochi, 2009). The latter, however, can only lead to reasonable estimates of changes in erosivity if the assumption of a stable relation between rainfall depth and erosivity holds true under changing boundary conditions.

The main objectives of this study are to use one of the very rare long-term (1937–2007) high resolution (≤5-min) data sets of ten stations in the central Ruhr area in Western Germany for a detailed analysis of: (i) long-term trends in summer rainfall erosivity (here: April–November), (ii) potential changes in frequency and/or magnitude of individual erosive rainfall events, and (iii) shifts in seasonality of rainfall erosivity.

Section snippets

Test area and data

The study area is located in the central Ruhr region in Western Germany ranging from the Lower Rhine Basin in its eastern part to the Westphalian Plain in its western part. In its South, it is bordered by the hills of the Rhenish Massif. The area is relatively flat with altitudes increasing from approximately 30 m a.s.l. in the West to 150 m a.s.l. in the East. In this densely populated and traditionally highly industrialized area, the local water authorities (Emschergenossenschaft and

Results

Following the data processing (Fig. 2) and the homogeneity tests, nine of the existing ten stations were chosen for trend analysis. SNHT indicates that summer rainfall at station Bochum DMT (Fig. 1, Table 1) is not homogeneous with a shift in the data set detected for the year 1975. This shift does not correspond with information regarding changes in measuring equipment and therefore might result from unknown changes in station set-up. Although no inhomogeneity was determined for summer

Discussion

In general, our results indicate a slight but significant increase in summer rainfall erosivity for the overall observation period (1937–2007) with a most pronounced increase during the last 35 years of observation (Fig. 4, Fig. 5). Due to the comprehensive data pre-processing and the multitude of statistical tests, we can be quite confident in these results. These trends are of most importance for erosion studies but they are also relevant for other hydrological studies dealing with extreme

Conclusion

A 71-year data set of high resolution (<5-min) rainfall data from nine stations in the central Ruhr area in Western Germany was used to analyze (i) trends in summer (April–November) erosivity, (ii) changes in the magnitude and frequency of individual erosive events, and (iii) changes in the seasonality of long-term mean daily erosivity.

For the period from 1937 to 2007, we observed a slight but significant increase in summer erosivity of 4.4% (in relation to the mean value) per decade. This

Acknowledgements

This research was carried out within the framework of the project “Effects of global climate change on the spatio-temporal variability in rainfall erosivity in North-Rhine-Westphalia, Germany” funded by the “Innovationsfonds – Anpassung an den Klimawandel in NRW”. Special thanks go to Dr. Andrea Hädicke for her support of the above mentioned project. The contributions to improve the English of the manuscript by Dr. B. Maxfield and the inputs by three anonymous reviewers should be also

References (39)

  • Canty, A., Ripley, B., 2012. R-package ‘boot’ (version 1.3-4). http://cran.r-project.org/web/packages/boot/index.html...
  • W.S. Cleveland et al.

    Local regression models

  • S.G. Coles

    An Introduction to Statistical Modelling of Extreme Values

    (2001)
  • Coles, S.G., 2011. R-package ‘ismev’: An introduction to statistical modeling of extreme values (version 1.37)....
  • A.C. Costa et al.

    Homogenization of climate data: review and new perspectives using geostatistics

    Math. Geosci.

    (2009)
  • M.J. Crawley

    The R Book

    (2009)
  • F. D’Asaro et al.

    Assessing changes in rainfall erosivity in Sicily during the twentieth century

    Hydrol. Proc.

    (2007)
  • Deutsches Institut für Normung, 2005. DIN 19708 - Bodenbeschaffenheit - Ermittlung der Erosionsgefährdung von Böden...
  • N. Diodato et al.

    Assessing and modelling changes in rainfall erosivity at different climate scales

    Earth Surf. Process. Landforms

    (2009)
  • Cited by (0)

    View full text