An average soil erosion rate for Europe: myth or reality?
The genealogy of this figure is of interest. [TABULAR DATA FOR TABLE 1 OMITTED] Crosson (1995) has already commented on the validity of the Pimentel et al. (1995a) erosion figure for the USA. This article focuses on erosion rates for Europe; it also explores the more general issues of the perpetuation of such figures in the literature and the validity of quoting an average erosion figure for a continental size area. The question needs to be asked: Is this the best available data, or are we witnessing scientific myth-making?
The propagation of a myth
The figure of 17 t [ha.sup.-1] [yr.sup.-1] (15,181 lb [acre.sup.-1]) in Pimentel et al. (1995a) has its origin in Bollinne (1982). However, by the time it reaches Pimentel it has travelled by a tortuous route and undergone several transformations (Table 1).
How was Bollinne's (1982) original figure transformed into that quoted by Pimentel et al. (1995a).
(1) In a review of several areas of Europe, Richter (1983) quotes Bollinne's (alleged) range of soil loss of 10 to 25 t [ha.sub.-1] [yr.sup.-1] (8,930 to 22,325 lb [acre.sup.-1] [yr.sup.-1]) obtained in central Belgium. The reference to "eight years of measurement, carried out in central Belgium" (Richter 1983), invites misinterpretation since no mention is made of the small experimental plots that Bollinne used to obtain this figure (see below).
(2) World Resources Institute (1986) reproduce Richter's figures without comment as part of a table of "Cropland erosion in selected countries." However, a note warns that some of the data in the table are point estimates based on a few measurements and extrapolated to reflect an annual average. Nonetheless, no indication is given of limitations of individual data items.
(3) In Lal et al. (1989) the figure from World Resources Institute (1986) is still 10 to 25 t [ha.sup.-1] [yr.sup.-1] (8,930 to 22,325 lb [acre.sup.-1] [yr.sup.-1]) but is applied now to Belgium rather than merely central Belgium. In the text of Lal et al. (1989) there is a generalized warning about comparisons of data from small plots because of unstandardized methodologies: "misinterpretation and erroneous conclusions are major worries when using such data" (p. 58).
(4) Barrow (1991 p. 211) reproduces the same table as Lal et al. (1989) with the 10 to 25 t [ha.sup.-1] [yr.sup.-1] figure for Belgium; a note with the table warns that data from different countries were obtained using different methodologies and the data "serves only as a general indication." Of the countries listed, only Belgium is in Europe.
(5) Pimentel et al. (1995a) transform 10 to 25 t [ha.sup.-]1 [yr.sup.-] (8,930 to 22,325 lb [acre.sup.-1] [yr.sup.-1]) into 17 t [ha.sup.-1] [yr.sup.-1] (15,181 lb [acre.sup.-1] [yr.sup.-1]), refer it to Europe, and credit it to Barrow (1991).
In Bollinne's (1982) original study of erosion rates in central Belgium, he uses four methods: measurement of rills on four fields over one winter; a six-year experimental study using 12 plots; measurement of the thickness of colluvium in dry valley bottoms; and measurement of colluvium in three enclosed hollows. For various reasons he rejects the data from the first three methods as either unsuitable or unreliable for broad generalization. Indeed, his caution extends to a refusal to present mean figures for the experimental plots. Average rates of erosion for the three hollows range from 12.8 to 15.6 t [ha.sup.-1] [yr.sup.-1] (114,304 to 139,308 lb [acre.sup.-1] [yr.sup.-1]). The figures are based on assumptions about the length of time since forest clearance; the contributing slope area; and the bulk density of the soil. Bollinne clearly states and discusses these assumptions.
Thus, Bollinne presents time-averaged erosion rates for three small areas in central Belgium over a period of about 150 years and is clearly aware of the problems of extrapolating the figures to a wider area. He discusses the issue of changes in erosion rates with the onset of modern mechanized farming, i.e., the validity of temporal extrapolation.
To be able to extrapolate spatially from the figures for the enclosed hollows Bollinne recognizes that assumptions have to be made about the representativeness of the sites in respect of land use, slope gradient, and length. These difficulties are discussed by Bollinne (1982).
Richter (1983), in his use of the figures in Bollinne's thesis, appears to ignore that part of the work the author considers most reliable and instead constructs a range of values from the plot experiments ('eight years measurement' according to Richter). It is these figures that enter the literature and are uncritically used by several authors in a process akin to the party game of Chinese Whispers. Finally, they are condensed into a single figure by Pimentel et al. (1995a).
Erosion rates in Europe
So how much erosion is really occurring in Europe? In the introduction to a recent overview of world erosion, Pimentel (1993 p. 2) asserts that, "soil loss rates in Europe range between 10 and 20 t/ha/yr," but he does not give a source. However, to obtain a sense that this figure, and that for the average erosion rate in Europe, is not adequate we need look no further than the only two chapters in Pimentel's book which deal with European countries, the United Kingdom and Poland. Arden-Clarke and Evans (1993 p.195) state that water erosion rates in lowland (i.e. arable) [TABULAR DATA FOR TABLE 2 OMITTED] Britain vary from 1 to 20 t ha-' [yr.sup.-1] (893 to 17,860 lb [acre.sup.-1] [yr.sup.-1]) but the latter are rare and localized events. They suggest that, 'most rates of erosion are less than 1 to 2 [m.sup.3]/ha/field/yr....' Wind erosion is less significant in Britain than erosion by water (Evans 1990; Evans 1996 p.31). Ryszkowski (1993 p. 222) estimates average rates of water erosion in Poland 0.52 t [ha.sup.-1] [yr.sup.-1] (464 lb [acre.sup.-1] [yr.sup.-1]), a figure based on river yields and an estimate of delivery ratio. These figures suggest that Pimentel's average erosion rate is about 15 times too high if applied to the two countries.
In addition, there are independent estimates of erosion rates from field monitoring schemes in Europe that suggest if we have to use average figures then 1 and 5 t [ha.sup.-1] [yr.sup.-1] (893 and 4,465 lb [acre.sup.-1][yr.sup.-1]) is a realistic estimate, but that there is considerable variability in space and time (Table 2).
The most comprehensive European monitoring scheme was that of the Soil Survey of England and Wales for seventeen localities in the years 1982-86 (Evans 1988; 1992). Data for the first three years are given in Table 3. About 700 [km.sup.2] (270 [mi.sup.2]) of farmland were included in the survey with fields affected by erosion accounting for 16.6 (6.41) in 1982, 40.5 (15.44) in 1983, and 9.8 [km.sup.2] (3.78 [mi.sup.2]) in 1984. Most major soil associations were sampled. Median volumes of soil eroded are for eroded fields, not the whole landscape. Values are relatively low.
In a survey of 86 fields over three winter periods, Govers (1991) reports a mean rill erosion rate of 3.6 t [ha.sup.-1] [yr.sup.-1] (3,215 lb [acre.sup.-1][yr.sup.-1]) for fields around Leuven, central Belgium.
Monitoring of about 36 [km.sup.2] (13.9 [mi.sup.2]) of farmland on the South Downs in southern England clearly shows year-to-year variations in erosion rates (Boardman 1992; Table 2). The wettest year of 1987 gave rise to an average rate about ten times higher than the driest year. Using data from this survey, the great spatial variation in erosion rates within the same area and for the same year can be shown [ILLUSTRATION FOR FIGURE 1 OMITTED].
Averaging. Most authors simply quote Richter's figure which is a range of about 10 to 25 t [ha.sup.-1] [yr.sup.-1] (9,000 to 22,000 lb [acre.sup.-1] [yr.sup.-1]). The transposition to a single figure of 17 t [ha.sup.-1] [yr.sup.-1] (15,181 lb [acre.sup.-1] [yr.sup.-1]) is done by Pimentel et al. (1995a). There is no explanation and it would seem unfortunate to lose the range of values originally given. However, the validity of Richter's version of the original is clearly open to question.
A more general point in relation to a suggested average value for erosion in Europe is that averages have usually been represented by the arithmetic mean. However, erosion events show a markedly left-skewed distribution with the mean therefore overestimating the central tendency. The median is a better descriptor of average erosion rates (Boardman 1988, 1990; Boardman et al. 1990; Evans 1988, 1990a).
Referencing. The usual scientific convention of referencing the source of information allows the reader to obtain further detail as to how the figure was obtained. For estimates of rates of soil erosion it is particularly important to know the methods used. To discover the basis for Pimentel's figure for the 'average rate of erosion in Europe,' required visits to two libraries to track down sources and to the [ILLUSTRATION FOR FIGURE 3 OMITTED] University of Liege to consult Bollinne's thesis; this should not be necessary.
Extrapolation. Bollinne used 12 experimental plots measuring 22.13 x 4m on a 6.5% hillslope over a six-year period (1974-79) at Sauveniere, Belgium (Bollinne 1982; Sinzot et al. 1989). He discusses at length the problem of extrapolation of these results to wider areas. For that reason he attempts to adapt the USLE to central Belgium rainfall conditions (Sinzot et al. 1989). The uncritical use of Bollinne's plot data results in the extrapolation of figures relating to an area of about 0.11 ha (0.27 acres) to represent either the total agricultural area or the arable area of Europe (Pimental et al. (1995a), are not clear on this point; the respective figures for 1992 are 126,813,000 (313,228,110) and 66,561,000 ha (164,405,670 acres) (European Community 1995).
The problem of extrapolation is illustrated by Figure 1 - a relatively small area with similar soils dominated by winter cereals. But where is the ideal site for experimental plots in order to obtain representative average erosion rates? Should plots be sited in an erodible part of the landscape or in parts where erosion is not occurring?
Evans (1995) compares erosion rates obtained from plot experiments, plot-based formulae and field measurements:
"In general, erosion measured on plots or estimated from plot-based formulae are considerably higher than those measured in farmers' fields, by a factor of between two and ten (p.123)."
This factor needs to be taken into account in interpreting the work of Bollinne and in any attempt to extrapolate his results to agricultural fields.
Comparability. The Pimentel et al. (1995a) estimate of 17 t [ha.sup.-1] [yr.sup.-1] (15,181 lb [acre.sup.-1] [yr.sup.-1]) for erosion in the USA is for a combination of wind and water. The figure for Europe is based on water erosion experiments alone (Bollinne 1982). It is therefore not comparable with the US figure. Because Pimentel et al. (1995a) do not apply their economic modeling to soil loss and productive capacity in Europe the distinction may be thought to be of little importance. However, uncritical reproduction and use of published figures is at the heart of this issue and therefore the point is worthy of note.
Table 4. Mean amounts eroded based on different choices of measurement area (Evans 1992) Amounts eroded [m.sup.3]/ha Shropshire Bedfordshire 1983 1984 1983 1984 Area within the field directly affected by erosion 69.41 40.99 33.88 20.00 Contributing area within field 16.22 3.75 1.04 0.32 Field 4.06 1.41 0.48 0.20 Transect 0.63 0.05 0.0081 0.0002
Average figures for regional erosion may be calculated in different ways. It is not clear if Pimentel et al. (1995a) intend their figure for Europe to apply to eroded fields or to the total landscape. Evans (1992) shows that erosion rates vary with the method of calculation (Table 4). Any "average rate" is highly dependent on the methods used and these must be clearly stated.
Use of field data. In a critical response to Pimentel et al. (1995a), Crosson (1995) suggests that the threat to global food supplies posed by erosion is grossly exaggerated by the authors. Pimental et al. (1995b) respond by claiming that Crosson's estimates are based on modeling whereas, 'we use data from field experiments of soil scientists for our assessment.' However, analysis of the derivation of the average figure for Europe shows that a set of field experiments by a careful scientist has been uncritically adapted, modified, and extrapolated and used for a purpose for which they were not intended. The transformation is achieved by a series of steps and a failure by several authors to check the basis of the figures they were quoting. Several warn of the problems of using or comparing erosion data from different sources but the warnings go unheeded and as a result data is recycled uncritically.
Perpetuation of a myth. Pimentel et al. were the last in a chain of misinterpretation; but scientific myths acquire a force of their own. In the year subsequent to publication, the Science Citation Index lists ten references to Pimentel et al. (1995a). One was the robust response of Crosson (1995). Other authors cite the article in support of the idea that erosion impacts on health and food supply (Westra 1996); aquatic systems (Pearce 1995); the productive capacity of soils (Mangel et al. 1996) and biodiversity and ecosystems (Harvey and Pimentel 1996). De Jong et al. (1995) cite it as showing the ecological benefits of soil conservation.
Some authors are more specific. Myers (1996) uses erosion rates and costs from Pimentel et al. (1995a), some of which are recycled from other sources, in support of his argument that vegetation and biodiversity protect soil from erosion. Walsh and Brown (1995) use the Pimentel at al. figure for the cost of soil erosion in Life Cycle Analysis of relative costs of conventional and organically grown cotton. The costs depend on assumed erosion rates. Doran et al. (1996) quote the Pimentel et al. (1995a) figures for off-site damages of erosion that are independent of their erosion rate. Harris (1996) refers to the Pimentel et al. (1995a) estimate of, "a yield loss of 20% from moderate erosion over a 20-year period" but recognizes that there are other estimates of yield loss. Harris (1996) suggests that although at present degradation is offset by fertilizer inputs, in future, resource and environmental factors will pose serious constraints on food supply.
The Pimental et al. figure has also been quoted by the national media in Britain as evidence of the seriousness of the erosion problem (Radford 1995, 1996). This analysis suggests that in the short period since publication Pimentel et al. (1995a) have been quoted largely in support of generalised arguments about the impact of soil erosion rather than specific usage being made of their rates and costs. However, the force of the article is clearly in its quantitative assessment of erosion and the costs of maintaining food supply. The data is used in support of the general argument about the seriousness of erosion; without the quantities the argument is significantly weakened. Arguments about the future sustainability of agriculture (cf. Harris 1996), clearly depend on reliable estimates of population, yield, and degradation rates.
The figure of about 17 t [ha.sup.-1] [yr.sup.-1] (15,181 lb [acre.sup.-1][yr.sup.-1])(Pimentel et al. 1995a) is of no value as it is a mean of a range of figures obtained from plot experiments which were rejected by the experimenter, Bollinne (1982), as an unsuitable base for spatial and temporal extrapolation. It would have been more acceptable to use Bollinne's preferred figures from the enclosed hollows but still quite unacceptable to extrapolate those to Europe. In the 1980s and 90s several field-based studies in Europe have addressed the issue of variation in erosion rates through time and space (Table 2). These are greatly to be preferred to plot-based studies as a method of providing regional average rates of erosion. Although they are not conclusive, and must be used with care, they indicate that mean rates are probably an order of magnitude lower on arable land in northern Europe than the figure used by Pimentel et al. (1995a). Even a cursory review of the results of these studies suggests that it is unjustified to use a single figure for the European continent.
The use of average rates for large areas distracts attention from the real issue: rates of erosion and associated on-farm and off-farm costs are high in some areas and not in others. They also vary temporally (e.g., Boardman 1992; Ryszkowski 1993), and are significantly affected by changes in land use and farming practice. Thus, the main scientific and political challenge is to gain acceptance for a targetted response, whereby resources are directed at areas and land uses where the risk is high rather than spread equally across the landscape. The use of an average figure for an area the size of a continent is not helpful, particularly when field measurements suggest that a value an order of magnitude less would be more appropriate. Winning the political battle for a correct appreciation of the threat of erosion, and achieving an appropriate response, will not be possible unless science uses figures that can be defended in a court of law, at least as the best available, rather than those that properly form the stuff of fairy tales.
Alstrom, K., and A. Bergman-Akerman. 1992. Contemporary soil erosion rates on arable land in southern Sweden. Geografiska Annalet 74A (2-3): 101-108.
Arden-Clarke, C., and R. Evans. 1993. Soil erosion and conservation in the United Kingdom. In: D. Pimentel (ed), World Soil Erosion and Conservation. Cambridge University Press, Cambridge, UK, pp. 192-215.
Barrow, C.J. 1991. Land degradation. Cambridge University Press, Cambridge, UK.
Boardman, J. 1988. Severe erosion on agricultural land in East Sussex, UK, October 1987. Soil Technology 1,333-348.
Boardman, J. 1990. Soil erosion on the South Downs: a review. In: J. Boardman, I.D.L. Foster, & J. Dearing (eds), Soil erosion on agricultural land, Wiley, Chichester, UK. pp. 87-105.
Boardman, J., R. Evans, D.T. Favis-Mortlock, and T.M. Harris. 1990. Climate change and soil erosion on agricultural land in England and Wales. Land Degradation and Rehabilitation 2(2): 95-106.
Boardman, J. 1992. Agriculture and erosion in Britain. Geography Review 6 (1), 15-19.
Boardman, J., and D.T. Favis-Mortlock, 1993. Simple methods of characterizing erosive rainfall with reference to the South Downs, southern England. In: S. Wicherek (ed), Farm land erosion: In temperate plains environment and hills. Elsevier. pp. 17-29.
Bollinne, A. 1982. Etude et prevision de l'erosion des sols limoneux cultives en Moyenne Belgique. These presentee pour l'obtention du grade de Docteur en Sciences Geographiques, Universite de Liege.
Crosson, P. 1995. Soil erosion estimates and costs. Science 269: 461-464.
De Jong, B.H.J., G. Montoya-Gomez, K. Nelson, L. Soto-Pinto, J. Taylor, and R. Tipper, 1995. Community forest management and carbon sequestration: A feasibility study from Chiapas, Mexico. Interciencia 20(6), 409-4 16.
Doran, J.W., M. Sarrantonio, and M.A. Liebig, 1996. Soil health and sustainability. Advances in Agronomy 56: 1-54.
European Community 1995. The agricultural situation of the European Union. 1994 Report. Office for the Official Publications of the European Communities, Luxembourg, 1995. p.432
Evans, R. 1988. Water erosion in England and Wales. Soil Survey and Land Research Centre, Silsoe, UK.
Evans, R. 1990a. Water erosion in British farmers' fields - some causes, impacts, predictions. Progress in Physical Geography 14(2), 201-219.
Evans, R. 1990b. Soils at risk of accelerated erosion in England and Wales. Soil Use and Management 6(3): 125-131.
Evans, R. 1992. Assessing erosion in England and Wales. Proceedings 7th ISCO Conference: People protecting their land. Sydney, Australia. pp. 82-91.
Evans, R. 1995. Some methods of directly assessing water erosion of cultivated land - a comparison of measurements made on plots and in fields. Progress in Physical Geography 19(1): 115-129.
Evans, R. 1996. Soil erosion and its impacts in England and Wales. Friends of the Earth, London.
Govers, G. 1991. Rill erosion on arable land in central Belgium: Rates, controls, and predictability. Catena 18, 133-155.
Harris, J.M. 1996. World agricultural futures: Regional sustainability and ecological limits. Ecological Economics 17(2), 95-115.
Harvey, C.A., and D. Pimentel. 1996. Effects of soil and wood depletion on biodiversity. Biodiversity and Conservation 5:1121-1130.
Lal, R., G.F. Hall, and F.P. Miller, 1989. Soil degradation: 1, Basic processes. Land Degradation and Rehabilitation 1, 51-69.
Ludwig, B., J. Boiffin, J. Chadoeuf, and A-V. Auzet. 1995. Hydrological structure and erosion damage caused by concentrated flow in cultivated catchments. Catena 25,227-252.
Mangel, M. et al. 1996. Principles for the conservation of wild living resources. Ecological Applications 6(2), 338-362.
Myers, N. 1996. Environmental services of biodiversity. Proceedings National Academy Science USA 93: 2764-2769.
Pimentel. D. 1993. Overview. In: D. Pimentel (ed) World Soil Erosion and Conservation. Cambridge University Press, Cambridge, UK, pp. 1-5.
Pearce, J.B. 1995. Urban development and marine and riparian habitat quality. Marine Pollution Bulletin 30(8): 496-499.
Pimentel, D., C. Harvey, P. Resosudarmo, K. Sinclair, D. Kurz, M. McNair, S. Crist, L. Shpritz, L. Fitton, R. Saffouri, and R. Blair. 1995a. Environmental and economic costs of soil erosion and conservation benefits. Science 267, 1117-1123.
Pimentel, D., C. Harvey, P. Resosudarmo, K. Sinclair, D. Kurz, M. McNair, S. Crist, L. Shpritz, L. Fitton, R. Saffouri, and R. Blair. 1995b. Soil erosion estimates and costs: Response science 269: 464-465.
Radford, T. 1995. Wearing the world away. The Guardian. March 9, 1995, London.
Radford, T. 1996. Population growth feeds a world crisis. The Guardian. April 13, 1996, London.
Richter, G. 1983. Aspects and problems of soil erosion hazard in the EEC countries. In: Prendergast, A.G. (ed), Soil erosion. Directorate General Agriculture, Commission of the European Communities. pp. 9-17.
Ryszkowski, L. 1993. Soil erosion and conservation in Poland. In: Pimentel, D. (ed), World soil erosion and conservation. Cambridge University Press. Cambridge. pp. 217-232.
Sinzot, A., A. Bollinne, A. Laurant, M. Erpicum, and A. Pissart. 1989. A contribution to the development of an erosivity index adapted to the prediction of erosion in Belgium. Earth Surface Processes and Landforms 14: 509-515.
Walsh, J.A.H., and M.S. Brown, 1995. Pricing environmental impacts: A tale of two T-shirts. Illahee 11,175-182.
Watson, A., and R. Evans. 1991. A comparison of estimates of soil erosion made in the field and from photographs. Soil and Tillage Research 19, 17-27.
Westra, L. 1996. Environmental integrity, racism, and health. The Science of the Total Environment 184, 57-66.
World Resources Institute 1986. An assessment of the resource base that supports the global economy. World Resources Institute. International Institute for Environment and Development.
John Boardman is a geomorphologist and Programme Leader for Land Degradation and Rehabilitation at the Environmental Change Unit, University of Oxford, 5 South Parks Road, Oxford OX1 3UB, UK. The author would like to thank David Favis-Mortlock and Bob Evans for critical comments on the article. He is indebted to Albert Pissart (University of Liege) for making available and discussing the thesis of his former student, Arthur Bollinne. This manuscript was submitted in December of 1996 and accepted in August of 1997.
|Printer friendly Cite/link Email Feedback|
|Publication:||Journal of Soil and Water Conservation|
|Date:||Mar 22, 1998|
|Previous Article:||Comparison of standard water quality sampling with simpler procedures.|
|Next Article:||Linking with agricultural input suppliers for technology transfer: the adoption of vertical tillage in Bolivia.|