Catena 50 (2003) 199 216 www.elsevier.

com/locate/catena

Ephemeral gully erosion in northwestern Spain

M. Valcarcel a, Ma . T. Taboada b, A. Paz b,*, J. Dafonte a
a

Escuela Politecnica Superior de Lugo, University of Santiago de Compostela, Lugo, Spain b Facultad de Ciencias, University of Corunna, Corunna, Spain

Received 20 October 2000; received in revised form 25 October 2001; accepted 10 December 2001

Abstract This study aimed to describe types of ephemeral gullies and to determine their origin, evolution and importance as sediment sources in A Coruna province (northwest Spain). Ephemeral gullies and/or rills have been measured in a representative sample of medium-textured soils, most prone to crusting, developed over basic schist. This sample consisted of 11 small sites, ranging from 0.63 to 7.34 ha. A case study of concentrated (rill + gully) erosion in a 0.47-ha catchment with coarsetextured soils developed over granite was also reported. The mean slope of the sites studied ranged from 6.1% to 16.8%. Main periods when soil surface was poorly covered were late spring (maize seedbeds) and late autumn early winter (grassland and winter cereal seedbeds). Case studies where fields were left bare in winter were also investigated. Soil incision and channel formation were observed even with relatively low rainfall intensities when the soil surface was sealed, but also after a single short intense rainfall event on recently tilled surfaces. Concentrated erosion took place mainly on seedbeds and newly tilled soils in late spring and by autumn or early winter, but gullies also appeared in other seasons when the soil surface was left bare. In most of the cases studied, ephemeral gully erosion caused significant soil losses, ranging between 2 and 5 m3/ha for a single season to locally, over 25 m3/ha. Gully development was significantly affected by agricultural operations, such as lineal elements often acting as initial axes of concentrated erosion. Main gullies tended to reappear at the same position. D 2003 Elsevier Science B.V. All rights reserved.
Keywords: Concentrated erosion; Ephemeral gully; Crusting; Tillage; Atlantic Spain

1. Introduction Regions along the Atlantic coast in northern and northwest Spain are characterized by humid, temperate climate. Rain intensities are moderate to low, like in other Atlantic
* Corresponding author. Tel.: +34-981-167000; fax: +34-981-167065. E-mail address: tucho@udc.es (A. Paz). 0341-8162/03/$ - see front matter D 2003 Elsevier Science B.V. All rights reserved. PII: S 0 3 4 1 - 8 1 6 2 ( 0 2 ) 0 0 1 3 9 - X

200

M. Valcarcel et al. / Catena 50 (2003) 199216

areas in Western Europe, extending from northern Portugal to the Scandinavian countries. Mean rainfall for the autonomous community of Galicia (northwest Spain) is in the range of 1400 1500 mm. Summers are often characterized by low total rainfall depths and dryness, even though thunderstorms with high-intensity rainfall are more frequent in this season (Font-Tullot, 1983). Thus, because of the water deficit in summer, the rain regime of Galicia presents, to some extent, transitional features between Atlantic and Mediterranean conditions. Rain erosion rates are expected to be moderate to high in a global perspective (Daz-Fierros and Daz de Bustamante, 1980; Gabriels, 2000). Soil degradation and erosion is a matter of concern in areas of Western Europe devoted to intensive agriculture. A number of field surveys demonstrated that concentrated flow erosion is widespread in different regions of the Loes belt, for example, north and northwest of France (Auzet et al., 1993; Ludwig et al., 1995, 1996), South Downs in England (Boardman, 1990), central Belgium (Govers, 1987, 1991; Poesen, 1989; Poesen et al., 1996; Nachtergaele and Poesen, 1999), the Limburg province in the Netherlands (Kwaad, 1991) and southwest Germany (Baade et al., 1993). Concentrated flow erosion was also described in other European regions with different agricultural systems and soil types, such as in the Scandinavian countries (Uhlen, 1986; Oygarden, 1996; Hasholt et al., 1997) or in the Lake Leman area (Vansteelant et al., 1997). More recent case studies for this type of erosion were also reported for two Iberian regions located in the transition zone between Atlantic and Mediterranean climate, southern Navarra (Casal et al., 1999) and northeast Portugal (Figueiredo de et al., 2000). These transitional regions are also characterized by low to moderate rain intensities, contrasting with typical Mediterranean environments where high intensity rains are more frequent. The importance of ephemeral gullies as sediment sources for the climatic conditions prevailing in Western Europe has also been stressed (Poesen, 1989; Poesen and Govers, 1990; Poesen et al., 1996). In Galicia, large, nonagricultural areas may be seriously affected by forest fires in dry summers. Thus, because of concern for soil degradation and water erosion after soil burning, earlier erosion research focused on forest and on shrubland (Daz-Fierros et al., 1982, 1990). However, runoff and erosion studies in agricultural land in northwest Spain have only been a recent matter of interest (Dafonte, 1999; Valcarcel, 1999). In fact, as in most of the Western Europe regions, attention has been given to agricultural soil erosion in Galicia, not because of concern for loss of soil fertility, but rather because of nutrients originating from the cultivated area causing water eutrophication. Starting in Spring 1997, we measured rill and ephemeral gully volumes in successive campaigns (Valcarcel, 1999). Despite the relative lack of information on soil erosion, large agricultural areas of a northwest Spain, particularly the Ordenes Basin (A Corun province), are affected by soil degradation and soil loss. In the intensive agricultural areas of northern Europe, most of the concentrated erosion surveys were conducted on catchments ranging from about 3 to 200 ha (Auzet et ` al., 1993; Ludwig et al., 1995, 1996; Vandaele and Poesen, 1995; Souchere et al., 1998). However, there are also areas in Europe where traditional agriculture is still important or where both traditional and intensive management systems are found side by side. Runoff generation and sediment production measurements have also been conducted at the small

 M. Valcarcel et al. / Catena 50 (2003) 199216

201

catchment (0.3 3.2 ha) scale (Oygarden, 1996) or even at the field level (Hasholt et al., 1997; Vansteelant et al., 1997). At this small scale, gully erosion in the uppermost part of a dry valley or along linear features in valley side or hillslopes has received little attention. In Galicia, traditional agricultural systems were characterized by the small size of fields and by a complicated system of terraces and border features separating the fields. For several centuries, thousands of kilometers of walls acting as terraces have been constructed at the property boundaries and they have been thought to be an important element in erosion control. In recent years, in some areas, properties have been redistributed, increasing the average field size and facilitating intensive farming practices. The extent to which agricultural land rearrangement measures may or may not increase erosion has not been investigated. The aim of this work was to present partial results on the contribution of ephemeral gullies to concentrated flow erosion. Both traditional and reallocated landscapes were investigated. We focused on the rates and extent of ephemeral gully erosion and on the description of the gully types at the small catchment and slope unit level in a period of two consecutive years.

2. Site description and methods The study area was located in northwest Spain, near the Atlantic coast, at a 30-km a. radius of A Corun Concentrated flow erosion was surveyed in 13 sites comprising only a, agricultural land. The study sites were situated around A Corun Laracha, Meson do Vento and Betanzos (Fig. 1). The altitude above sea level ranged from approximately 80 to 480 m. The investigation extended over two consecutive years, 1997 1998 and 1998 1999, reported observations starting April 1997.

Fig. 1. Location of the study sites.

202

M. Valcarcel et al. / Catena 50 (2003) 199216

2.1. Regional context: climate and soils In the study area, the mean annual rainfall was about 1000 1500 mm, lower than the average for Galicia, estimated in the range of 1400 1500 mm (Font-Tullot, 1983). Rainfall measurements for three rain gauges, A Coruna, Mabegondo and Ordenes, representative of the study area, are summarised in Table 1 for 1997 1998 and 1998 a 1999. Mean annual rainfall in A Corun was 1017 mm, this value increases with altitude above sea level from coast to inland zones. The between-year variations in rainfall were smaller than the variations between lowland and highland. However, the seasonal rainfall for different years showed important variations. In Mabegondo, the average rainfall at intensities higher than 10 mm/h was about 7% of total precipitation (Valcarcel, 1999). Twenty-three events at intensities higher than 10 mm/ h were recorded in the 2-year study period. Most intense rains occurred in spring; however, more short intense rainfalls were recorded in Spring 1997 than in Spring 1998. The highest rain depth for a 24-h period was 52.1 mm on June 1, 1998, corresponding to a return period of about 50 years. The study sites are located in two contrasting geologic and pedologic conditions. The two different soil parent materials characterizing the area are basic metamorphic rocks, mainly basic schist, and granite bedrock. Metamorphic basic schist belongs to the lithotectonic unit of the Ordenes Basin, the most common basic rock formation found in the northwest of the Iberian Peninsula. Soils on basic schist are relatively deep medium-textured, mainly loam and loamy silt, according to USDA criteria, and rather fertile; this is attributed to an important alteration produced by weathering. Soils on granite are shallow, coarse-textured, mainly sandy loam and loamy sand and less fertile. Because of leaching losses induced by high precipitation, soils developed from both parent materials are acidic, but generally pH is much lower on granite soils, as expected due to the few basic ions in this acidic rock. Soils on both parent materials are rich in organic matter, but comparatively more organic matter is accumulated on the surface of soils on granite. Main soil types are classified as Umbrisol and Cambisol, according to the World Reference Base for Soil Resources (FAO-ISRIC, 1994). As shown in Table 2, basic schist and granite soils were unequally represented in the sample. Most of the study sites were located in the soils on schist of the Ordenes Basin,

Table 1 Seasonal and yearly rainfall (mm) for the study period April 1, 1997 March 31, 1999 Period Coruna 1997 1998 April 1 June 30 July 1 September 30 October 1 December 31 January 1 March 31 April 1 March 31 279.3 55.4 479.2 147.1 961.0 1998 1999 326.8 184.0 243.8 443.0 1197.7 Mabegondo 1997 1998 329.0 78.8 475.5 138.4 1021.7 1998 1999 435.4 122.3 214.1 365.1 1136.9 Ordenes 1997 1998 344.2 113.5 717.9 217.2 1392.8 1998 1999 500.3 220.2 269.2 428.1 1417.5

 M. Valcarcel et al. / Catena 50 (2003) 199216 Table 2 Main characteristics of study sites where concentrated flow erosion was measured No. Site name and code Area (ha) Mean slope (%) 6.1 15.3 13.1 12.3 11.9 8.3 9.2 7.1 7.8 9.3 7.0 Percentage of area in slope class < 5% 5 10% >10% No. of fields

203

Soil texture

Schist 1 2 3 4 5 6 7 8 9 10 11

area of the Ordenes Basin Mabegondo (M1) 2.54 Linares (LI) 1.48 Mabegondo (M2) 0.63 Laracha (LA) 0.84 Mabegondo (M3) 5.39 Mabegondo (M4) 1.28 Abelar (AB) 2.89 Mabegondo (M5) 5.72 Mabegondo (M6) 7.34 Mabegondo (M7) 5.37 Mabegondo (M8) 6.26

30.7 1.6 10.6 1.7 9.7 10.9 18.4 15.1 23.9 23.9

69.3 7.4 22.4 29.2 33.8 80.0 67.7 70.4 64.6 45.5 45.5

91.0 77.6 63.1 64.5 10.3 21.4 11.2 20.4 30.6 30.6

3 7 1 2 3 2 1 7 3 4 2

Silt-loam Loam Silt-loam Sandy loam Silt-loam Silt-loam Loam Silt-loam Silt-loam Silt-loam Silt-loam

Granite area near Coruna 12 Revoltas (RE) 13 Rialta (RI)

a

0.47 0.07

7.8 16.8

5.4

94.6

100

3 1

Sandy loama Sandy loama

Soils with granitic bedrock.

where surface crusting had been frequently observed (Taboada et al., 1999). The sample for concentrated flow erosion surveying on medium-textured soils of the Ordenes Basin comprised 11 sites. Two sites were on soils with granite underneath. (Table 2). Most of the topsoil horizons in the study sites of the Ordenes Basin were silt-loam with high silt contents (55 65%) and moderate clay contents (10 20%), but soils with loam or sandy loam texture were also investigated in this area. Soil textures in the studied sites on granite were sandy loam with more than 60% sand content. 2.2. Selection of study sites In Galicia, as a consequence of land partitioning with time, the average agricultural property size has been estimated at about 4 5 ha, making the agricultural production system very unique and family oriented. Even after property redistribution, field sizes remain much smaller than those in intensive agriculture regions of Spain and northern Europe. Representative landscapes for traditional and intensive agriculture systems have been the primary criteria for the selection of sites. Thus, concentrated flow erosion was surveyed on two major types of landscapes, old agricultural landscapes with traditional land uses characterized by terraces and extremely small fields, and sites with increased field size originated by land redistribution, which are more suitable to mechanized agriculture. Sites in the Mabegondo zone (Fig. 1) were also selected because they were part of ongoing research programs or because previous erosion events were known. Despite the complexity of the agricultural landscape, all 13 experimental sites were selected as areas limited by permanent features, i.e. roads, ditches and water divides and/ or temporary barriers such as dead furrows and other linear elements of agricultural

204

M. Valcarcel et al. / Catena 50 (2003) 199216

origin. Thus, each site was well isolated by border features regarding surface runoff generation. Morphologically, the study sites were small catchments or hillslopes. Sites where a small catchment was well defined and also some of the hillslopes were hydrological isolated units as runoff water concentrates and conveys to one single outlet. The main concentration line was either a segment of the valley floor or a linear feature of agricultural origin (Auzet et al., 1993). However, in other hillslopes, runoff was diverted into more than one outlet. In this case, the study site was divided into slope units, defined by Hasholt et al. (1997) as small independent source areas with regard to surface runoff. The surface area of the study sites extended from only 0.07 to 7.34 ha (Table 2). In turn, most of the sites, and particularly those on old landscapes, were not homogeneous in terms of crops and tillage as they were divided into small fields with different land uses. The number of individual fields (or plots) within each site ranged between 1 and 7 (Table 2). Altogether, 39 small fields were surveyed; their size varied between 0.025 and 4.68 ha and the six smallest fields were less than 0.2 ha in surface area. This range of sizes is a representative of fields in A Corun province. 2.3. Field survey Data on topography and land use were collected from field surveys. Topographical surveys were conducted using an Abney level from which a detailed topographic map was elaborated for each study site. DTM were also elaborated using the software package PCRaster (Karssenberg, 1996). Physical parameters shown in Table 2 (mean slope, percent of slope for selected classes) were obtained from these maps. Most of the sites were characterized by gentle slope, maximum slope values generally being less than 15%. Nevertheless, mean slopes of two study sites were steeper than 15%. Field observations were made after each important rainfall event and were particularly frequent when the soil surface was uncovered. Following the methodology previously described by Auzet et al. (1993) and Ludwig et al. (1995), field observations included land use data monitoring and assessment of soil surface degradation state according to the criteria first proposed by Boiffin et al. (1988). Rill and gully systems were monitored. The position of each channel was located by Abney level in each survey, allowing representation on the topographic map of each site. Sediment production was evaluated by measuring channel volume (Govers, 1987; Ludwig et al., 1995). Width and depth of rills and ephemeral gullies were measured every 2 5 m and at specific points, where the section changed abruptly. Eroded volumes were calculated from cross-sections and length of channel segments. The threshold between rills and gullies was considered to be 1 ft2 (929 cm2) according to criteria proposed by Hauge (1977) and discussed by Poesen et al. (1996). 2.4. Laboratory methods Soil organic matter content and soil texture were determined by standard laboratory methods as quoted by Valcarcel (1999). Bulk density of topsoil and subsoil was also measured in some of the study sites following concentrated flow erosion episodes.

 M. Valcarcel et al. / Catena 50 (2003) 199216

205

3. Results and discussion Main dominant cultivations in the study sites were maize and grassland, but some small fields were used for winter cereals, potatoes, orchards and rape; in addition, sometimes, fields were left fallow during winter after maize. Rotations in the Ordenes Basin area were maize fallow, grassland maize, maize winter cereals. In the granite area, potatoes followed rape or winter cereals. Winter fallow was also observed, both in old landscapes and after land consolidation. In all of the 13 study sites, soil losses by concentrated flow erosion were documented during at least one survey date, over the 2-year study period. Cumulative erosion volumes and rates due to ephemeral gully incision, during this period varied in these sites between 6.02 10 2 and 55.60 m3/ha. Fig. 2 shows concentrate flow erosion rates for 10 out of 11 study sites of the Ordenes Basin. Catchment coded M7 was kept out of this figure because of the lowest concentrated flow erosion rate of 6.02 10 2 m3/ha cumulative, taken into account a mean bulk density of 1500 kg/m3 for the eroded horizons; mean annual rates for the 2-year study period varied between 5.01 10 3 and 4.17 kg/m2/year. The most important types of arable land where concentrated flow erosion was observed in the 11 surveyed sites of the Ordenes Basin were: tilled surfaces, prepared for sowing maize in later spring. tilled surfaces, prepared for sowing winter cereals or prairies in autumn. However, the occurrence of soil incision in land left fallow after maize all over the winter (e.g. sites M1 and AB) has not been significant, even if overland flow was frequently observed in such soil surfaces. This may be attributed to soil surface compaction during harvesting. In the two study sites of the granite area, concentrated flow erosion was caused by one single summer rainstorm in August 1997. The soil surface of both sites was freshly tilled just before the rainstorm. Previous crop was potato. Soil loss rates were as high as 49.96 and 31.71 m3/ha for site numbers 12 (Revoltas) and 13 (Rialta), respectively.

Fig. 2. Temporal evolution of concentrated (rill + gully) erosion in 10 sites of the Ordenes Basin.

206

M. Valcarcel et al. / Catena 50 (2003) 199216

Medium-textured soils of the Ordenes Basin have been found to be prone to crusting. Sedimentary crusts with a very low saturated conductivity ( < 5 mm/h) develop from freshly tilled surfaces after cumulative precipitation of about 150 200 mm. Soil surface kinetics was also found to depend on organic matter content and initial roughness of the tilled surface (Taboada et al., 1999). This means that after a structural crust has developed, the infiltration capacity during heavy rains is an order of magnitude lower than peak rain intensity. Therefore, runoff production by overland flow was likely to be the dominant process causing incision and concentrated flow erosion. Fig. 2 also shows important differences in concentrated flow erosion rates between the two study years. This was due, both to the temporal distribution and intensity of rainfall and to the proportion of soil covered by crops in a given season, which was imposed by the rotation schemes. For example, there were clearly important differences in concentrated flow erosion rates between later Spring 1997, and the same period of 1998. The most important soil losses for individual campaigns conducted during the 2-year study period were monitored in later Spring 1997. The volume of soil lost in study site numbers 1, 2, 3, 4 and 5 (i.e. Mabegondo, M1, M2, M3, Linares (LI) and Laracha (LA)) was in the range between 10.79 and 55.59 m3/ha. Unusually heavy rainfall events occurred, recurrently, on May 25, May 26, May 30 and June 2, 1997. Daily rainfall amounts of 35.8, 18.7, 18.9 and 24.4 mm, respectively, were recorded in Mabegondo during these dates. These four events, with a precipitation total of 97.8 mm caused heavy erosion in the seedbed of maize fields, already degraded to some extent by about 50 mm previous rainfall. Subsequently, surface crusting and channel erosion could be well documented in 1997. During 1998, maize was also sown in the same sites (numbers 1 to 5); however, by the end of June 1998, less than 10% of the surface had a structural crust, and consequently total soil erosion, including sheet erosion was not significant. Even if cumulative rainfall between April 1 and June 30 was higher during 1998, concentrated flow erosion in that period was negligible. There were also important differences between ploughed soils and seedbeds. No significant concentrated flow erosion was found with high amounts of rainfall after mouldboard plough in maize fields during early Spring 1997 and 1998 (e.g. sites M1 and M2 at Mabegondo). In this case, the absence of runoff generation may be attributed to the important storage capacity in depressions at the soil surface and was associated with the high roughness produced by ploughing (Kamphorst et al., 2000). However, because of the low surface roughness, autumn-tilled surfaces, left bare all over the winter (e.g. site M5 at Mabegondo), produced high rates of concentrated flow erosion as shown by measurements at the end of the winter. Grassland and winter cereal sowed in autumn originated surfaces with low roughness values, also prone to crusting where runoff was frequent. Examples were sites M6 and M8 at Mabegondo (Fig. 2). However, once soil cover reached its maximum, grassland prevented totally concentrated flow erosion. Because a temporal prairie protects the soil surface for a length of 3 or 4 years, concentrated flow erosion risk was highly reduced when rotations included grassland and permanent pastures. Thus, conventional tillage practices for seedbed preparation, both during spring and autumn, have most enhanced concentrated flow erosion. In contrast, the maintenance of grassland continuous cover completely prevented their formation.

 M. Valcarcel et al. / Catena 50 (2003) 199216

207

3.1. Patterns of gullying Ephemeral gullies were found in 6 out of 11 surveyed sites on the Ordenes Basin area and in one out of two sites on the granitic area (Table 3). Two main patterns of gullying, matching those previously described by Ludwig et al. (1995, 1996) and Poesen et al. (1996), were found: (i) gullies located on topographical features, such as thalwegs in the valley bottom of small-sized catchments and (ii) gullies associated with linear man made permanent or temporal agricultural features, mainly depression lines in hillslopes. Based on field observations, examples are next described. An example of gullying associated with agricultural lineal depressions, thus characterized by a nonuniform spatial distribution, is shown in Fig. 3. This ephemeral gully was located at the boundary of a hillslope, limited by a road (site M5 at Mabegondo). The slope length promoting concentrated flow was 325 m long. It was first observed by the end of October 1997; it developed further and it was erased in early Spring 1998. However, it reappeared at the field border already in later Spring 1998 and also one more time during Autumn 1998 Winter 1999. The contribution of the main channel (gully or rill + gully, depending on the season) to concentrated flow erosion in this site was always higher than 90%. This result points at the relative importance of gullying on concentrated soil erosion and for conveying water from the crusted surface to the outlet. An aerial view showing ephemeral gullying due to flow concentration into a topographically predetermined depression at the valley bottom of a small catchment (M1 site at Mabegondo) is shown in Fig. 4. This pattern of gully was also observed, but not frequently because zero-order catchments were poorly represented in the sample due to the small dimensions of the study sites. The site was well isolated all around the perimeter and

Table 3 Soil losses by ephemeral gully erosion and gully characteristics during the period 1997 1999 No. Site name and code Survey dates Ephemeral gully position in valley Bottom Bottom + side Bottom + side Bottom Bottom + side Side Side Side Side Side Losses by gullying (m3/ha) 0.72 3.81 2.46 26.14 2.94 3.94 5.28 0.17 5.07 1.31 Gully/rill + gully(%) Mean crosssection (m2) Width depth ratio 6.57 11.97 4.24 2.24 2.97 3.35 2.44 2.14 3.44 1.64

Schist of the Ordenes Basin area 1 Mabegondo (M1) May 1997 June 1997 2 Linares (LI) June 1997 4 Laracha (LA) June 1997 5 Mabegondo (M3) June 1997 8 Mabegondo (M5) November 1997 January 1998 May 1998 11 Mabegondo (M8) July 1998 January 1999 Granite area near Coruna 12 Revoltas (RE) August 1997

10.78 35.26 5.68 61.63 26.04 65.97 76.11 18.42 22.46 43.28

0.158 0.179 0.149 0.168 0.131 0.156 0.149 0.131 0.165 0.261

49.96

43.27

0.160

4.77

Volumetric losses by gullying are for surface units.

208

M. Valcarcel et al. / Catena 50 (2003) 199216

Fig. 3. Ephemeral gully in valley side, along a dead furrow associated with a field border.

runoff water generated upslope was conveyed and concentrated into a topographical depression at the valley bottom. An example of a dendritic rill pattern along rows and wheel tracks, which gradually converge into a gully along a micro-topographical depression at the field border of a hillslope is shown in Fig. 5. This ephemeral gully system was observed in later Spring 1998 in site M8 at Mabegondo. Again, this gully pattern is located in a pre-existing lineal structure initiated by tillage.

 M. Valcarcel et al. / Catena 50 (2003) 199216

209

Fig. 4. Aerial photograph showing ephemeral gullying in valley bottom.

In traditional agricultural landscapes with small fields separated by terrace walls, such as in Linares (LI) and Laracha (LA), a particularly adapted ephemeral gully pattern was found. In these cases, overland flow in the fields above the terrace wall was observed to produce rill interrill erosion, whereas gully erosion occurred in fields below the wall. Ephemeral gullying started at the most upslope part of the second field. Because wall-sides between neighbour fields located along the slope remained stable and firm, no bank erosion was produced. An example of parallel rill patterns, which was observed in site M5 at Mabegondo, is a gentle hillslope. This parallel network showed no convergence into a gully. Similar rill patterns had been also previously described in other regions (Govers, 1987; Ludwig et al., 1995; Casal et al., 1999). In this case, the channel network was imposed by very close wheel tracks separated by a regular distance of 1.80 m (Valcarcel 1999). In summary, most of the surveyed ephemeral gullies appeared to be governed by overland flow as a mere deepening of previous linear features, mainly in valley side topographic location. In some cases, gullies were promoted by convergent concentrated flow in drainage lines of small zero-order catchments, i.e. in valley-bottom. Singular gullies may appear in terraced, old agricultural landscape. The importance of the spatial position of runoff contributing areas on concentrated flow erosion in areas of intensive agriculture has frequently been stressed (Ludwig et al., 1995, 1996). The possibility of identifying runoff collectors in order to analyze the spatial variability of erosion at the small spatial scale used in this study was also investigated. A case study is presented in Fig. 6, where both DTM and a schematic representation of concentrated flow pathways and ephemeral gully location for study site M1 (Mabegondo),

210

M. Valcarcel et al. / Catena 50 (2003) 199216

Fig. 5. Gully along a field border as a result of downward convergence of a dendritic rill network.

at the end May 1997, are shown. It is apparent when comparing DTM and sketch that the natural way of conveying excess water has been strongly modified by agricultural land use. The small catchment comprises three fields, separated from each other by two unpaved roads. The detailed DTM shows that the first two fields located upslope are concave, whereas the field at the bottom of the catchment has a topographical depression delineating a thalweg. This example confirms that overland flow from areas upstream causes ephemeral gullying in a different field. In other words, it was illustrated how runoff produced in a place may cause concentrated flow erosion in another place. The case study also shows how local infrastructures and the agricultural framework influenced gully development in a small catchment. However, further work is needed to test the hypothesis of the relationship between concentrated flow erosion and hydrographical structure at the study scale. Thus, depending on rain intensities, on soil erodibility and on agricultural factors governing the water flow pattern, ephemeral gullies developed at small surface scale.

 M. Valcarcel et al. / Catena 50 (2003) 199216

211

Fig. 6. DTM and schematic representation showing main lines of flow concentration and gully erosion during late spring events in a site matching a small zero-order catchment. (Contours are in meters above sea level).

Spatial variability of ephemeral gully erosion was to a large extent dependent on both topographical and anthropogenic factors. Agricultural operations significantly affected gully development, as lineal features of the landscape often act as initial axes of erosion. 3.2. Relative contribution of ephemeral gully to sediment production Ephemeral gully volumes and main characteristics are shown in Table 3. Taking into account all the surveyed surfaces of the Ordenes Basin, the relative contribution of gully erosion to concentrated flow sediment production varied between 0% (four sites without gullying) and 76%. That percentage varied also throughout the year and between erosion events. In this area, ephemeral gully erosion caused significant soil losses. In most of the sites, where gullying was observed, the range of soil loss by this type of concentrated flow

212

M. Valcarcel et al. / Catena 50 (2003) 199216

erosion during the different measurement epochs was between 2 and 5 m3/ha. Minimum and maximum values in these six sites with gully erosion were 0.17 and 26.14 m3/ha, respectively. Taking into account the whole surveyed surface in the Ordenes Basin, rates of ephemeral gully erosion and concentrated flow erosion were calculated as shown in Table 4. The total surveyed surface for concentrated flow erosion was 36.8 ha. Mean cumulative soil losses computed for this surface, over the 2-year study period, due to ephemeral gully erosion were 2.12 m3/ha. Concentrated flow erosion by rills plus ephemeral gullies was in this surface, 5.89 m3/ha for the 2-year period. This means that ephemeral gully erosion was on average 26.4% of concentrated sediment production during the 2-year study period. Mean ephemeral gully erosion rate of 1.06 m3/ha/year matches well with results obtained in other regions with humid climate (Vandaele and Poesen, 1995; Poesen et al., 1996), even if it was somewhat lower than rates measured in areas of intensive agriculture in northern Europe. However, this rate is much lower than values found in areas with Mediterranean climates (Poesen et al., 1996; Alba de et al., 1998). Also, both concentrated flow erosion rates and the relative contribution of rill and gully volumes in the Ordenes Basin were in agreement with data reported for other regions with Atlantic climate. Gully erosion rates in the medium-textured soils of the Ordenes Basin are low or moderate in a global perspective. Critical periods for ephemeral gully formation were late spring and autumn early winter, as for concentrated flow erosion. Results also indicated the recurrence of ephemeral gully erosion. Ephemeral gullies in sites M5 and M8 at Mabegondo showed a trend to reappear at the same position. In site M5, three recurrent channel incision episodes were observed in the period of November 1997 January 1999. They were strong enough to originate ephemeral gully development. Notice the important total soil loss rates in the winter period of November 1997 January 1998 (5.28 m3/ha), when the highest relative contribution of ephemeral gully erosion to soil losses by concentrated erosion was recorded. Important soil losses associated with gullying were also more than one time observed in the site M8 at Mabegondo since 1998. An ephemeral gully was first observed in later Spring 1998 under maize, while this site was before used for grassland. The soil surface was prepared as a seedbed for prairie in Autumn 1998 and in January 1999 an ephemeral gully was again incised in the same position. In M8, the rate of soil loss was smaller in the winter period than in the later spring one, but the contribution of ephemeral gullying to total erosion was about double as high in winter than in later spring. Concentrated flow erosion measurements made more than one time within one erosive period allowed assessing the evolution of the relative contribution of ephemeral gullies as a function of cumulative rainfall. Results are shown in Table 3 for site M1 during later Spring 1997 and for site M5 during Autumn 1998 Winter 1999. They clearly show that
Table 4 Average erosion losses by rills and gullies in the Ordenes Basin sample during the period 1997 1999 Surveyed surface (ha) 36.8 Rill + gully (m3/ha) 8.01 Rill erosion (m3/ha) 5.89 Gully erosion (m3/ha) 2.12 Gully/rill + gully (%) 26.4

 M. Valcarcel et al. / Catena 50 (2003) 199216

213

the relative contribution of ephemeral gullies to total concentrated soil losses was increasing as the rainfall increased. The rise in the importance of ephemeral gully soil losses as a function of the number of erosive rainfall events during a single season may be partly attributed to the increased degradation of structure with time. The observed gully pattern in the granite area near Coruna, which is also shown in Table 3, will be discussed next. As before mentioned, in August 1997 a local heavy rainstorm with a total precipitation total of 29.8 mm caused channel incision in arable land, tilled with rotative implements after potato harvest in the outskirts of Coruna. The surface area of this field was 0.437 ha. Though there were scarce dimensions for producing overland flow, soil loss rates were about 50 m3/ha. Ephemeral gullies represented 43.27% of total soil losses. This result is more striking, if it is taking into account that previous to the rainstorm causing concentrated flow soil losses, the topsoil was not degraded by crusting, and the saturated hydraulic conductivity of coarse-textured soils on granite is very high. However, the sandy loam soil is thought to be very prone to develop hydrophobia when dry, which might explain the low water infiltrability during the rainstorm and the formation of a sheetwash regime on the dry topsoil. Excess water was concentrated along linear elements located very closely at distances of about 1.2 m and finally originating devastating gullying following wheeltracks. 3.3. Main characteristics of the gully systems Mean ephemeral gully widths, depths and cross-sections are shown in Table 3. Crosssections oscillated between 0.13 and 0.26 m2. Average values of width depth ratio were in the range between 1.63 and 11.97. Broad gullies had been considered to be more harmful for water quality because of relatively higher losses of topsoil rich in fertilizers and organic matter, which may cause nonpoint source loads in streams. In general, a trend was observed for higher width depth ratios by heavy rains in spring and smaller ratios in the winter season. The higher width depth ratio was recorded in the thalweg gully developed near the outlet of catchment M1 at Mabegondo. This was due to a significant higher erodibility in topsoil with regard to subsoil because of the presence of a plow pan. Mean bulk density was 1380 kg/m3 in the topsoil versus 1650 kg/m3 in the subsoil. Consequently, the surface horizon of about 20 cm in depth was eroded, but not the subsoil as shown in Fig. 7. Channel sections near the outlet were more than 1.6 m wide, whereas they were frequently 15 20 cm deep. Table 2 also shows that the smallest width depth ratios were recorded for site LI, near Laracha; in this case, the ephemeral gully bed was frequently located on the bedrock and gully depths of the order of 50 to 60 cm were measured. Ephemeral gullies formed by incision along linear elements generally showed large sections in zones with high slope, so that a gradual decrease from the maximum crosssection, both toward the head and downstream, occurred. This variation from head-cut to outlet of the gully may be attributed to the small flow rates at the upstream and saturation of the transport capacity downslope, where sedimentation initiates (Casal et al., 1999). Frequently, management systems in old agricultural landscapes included a number of measures for runoff control. Concentrated flow may be routed away from agricultural

214

M. Valcarcel et al. / Catena 50 (2003) 199216

Fig. 7. Detail of an ephemeral gully over a plow pan.

lineal features such as furrows into concrete channels along roads. The effect would be subdividing the runoff contributing area, thus concentrating flow volume and avoiding water from flowing in uncontrolled manner on agricultural land. Runoff routing is an elemental measure that probably would have reduced concentrated flow erosion, both on small catchments (site M1) and hillslopes (site M5) in the study conditions. Contouring, grassed waterways and tillage direction management would also reduce overland flow concentration and minimize erosion.

4. Conclusions In the medium-textured soils of the Ordenes Basin, occurrence of concentrate flow erosion was related to surface crusting. Human impact is demonstrated through variations

 M. Valcarcel et al. / Catena 50 (2003) 199216

215

caused by crop rotation and tillage procedures. In a representative sample of cultivated land of the Ordenes Basin, mean cumulative concentrated flow erosion rate over a 2-year period was 8.01 m3/ha. On average, ephemeral gully erosion contributed 26.4% (i.e. 1.06 m3/ha/year) to the concentrated flow erosion. Ephemeral gully caused by a heavy summer rainstorm was described also in coarse-textured soil on granite. Concentrated erosion may transport large amounts of sediment to streams, unless buffer zones between the eroded surface and the permanent watercourses were present. Conventional tillage practices and seedbed preparation enhanced concentrated flow erosion and gully occurrence, whereas the maintenance of vegetation cover completely prevented soil surface incision and channel formation. The study of the development of the gully system through time showed that main gullies tend to reappear at the same position. Acknowledgements This paper has been produced with the support of research projects HID96-1085-C0201 funded by CICYT and PGIDT99MA10303 funded by Xunta de Galicia, respectively. We thank two referees for their annotations and comments. References
Alba de, S., Benito, G., Perez-Gonzalez, A., 1998. Erosion de suelo en episodios de lluvia de elevada intensidad versus episodios de moderada y baja intensidad y elevada frecuencia en ambientes semiaridos. In: Gomez ola. Universitat de Ortz, A., Salvador, F. (Eds.), Investigaciones Recientes de la Geomorfologa Espan Barcelona, Barcelona, Spain, pp. 483 492. Auzet, A.V., Boiffin, J., Papy, F., Ludwig, B., Maucorps, J., 1993. Rill erosion as a function of the characteristics of cultivated catchments in the north of France. Catena 20, 41 62. Baade, J., Bersch, D., Mausbacher, R., Schurkraft, G., 1993. Sediment yield and sediment retention in a small loess-covered catchment in SW-Germany. Z. Geomorphol., Suppl.bd 92, 201 216. Boardman, J., 1990. Soil erosion in the South Downs: a review. In: Boardman, J., Foster, D.A., Dearing, J.A. (Eds.), Soil Erosion on Agricultural Land. Wiley, Chichester, UK, pp. 87 105. ` Boiffin, J., Papy, F., Eimberk, M., 1988. Influence des systemes de culture sur les risques derosion par ruisselle ment concentre: I. Analyse des conditions de declanchement de lerosion. Agronomie 8, 663 673. Casal, J., Lopez, J.J., Giraldez, J.V., 1999. Ephemeral gully erosion in southern Navarra (Spain). Catena 36, 65 84. Dafonte, J., 1999. Procesos hidrologicos superficiales en cuencas agrcolas. PhD Thesis. EPS de Lugo, Univ. of Santiago de Compostela, Spain, p. 190. Daz-Fierros, F., Daz de Bustamante, J., 1980. Erosividad potencial de las lluvias durante los anos 1978 79 y 1979 80. In: de Pablos, B. (Ed.), Avances de la Investigacion en Bioclimatologa. CSIC, Madrid, pp. 305 314. Daz-Fierros, F., Gil, F., Cabaneiro, A., Carballas, T., Leiros de la Pena, M.C., Villar, M.C., 1982. Efectos erosivos de los incendios forestales en suelos de Galicia. An. Edafol. Agrobiol. 41 (3 4), 627 639. Daz-Fierros, F., Benito, E., Vega, J.A., Castelao, A., Soto, B., Perez, R., Taboada, T., 1990. Solute loss and soil erosion in burnt soil from Galicia (NW Spain). In: Goldammer, J.G., Jenkins, M.J. (Eds.), Fire in Ecosystem Dynamics. Mediterranean and Northern Perspectives. SPS Academic Publishing, Amsterdam, The Netherlands, pp. 103 116. FAO-ISRIC., 1994. World Reference Base for Soil Resources. Draft. Rome, Italy/Wageningen, The Netherlands, p. 161. Figueiredo de, T., Poesen, J., Vandekerckhove, L., Oostwoud-Wijdenes, D., Araujo, J., 2000. Gully erosion under global change. International Symposium on Gully Erosion under Global change. Leuven, Belgium, 16 19 April 2000. Book of Abstracts. K.U. Leuven, Belgium, p. 49. http://www.kuleuven.ac.be/facdep/geo/fgk/leg/ pages/espgeom.htm.

216

M. Valcarcel et al. / Catena 50 (2003) 199216

Font-Tullot, I., 1983. Climatologa de Espana y Portugal. INM, Madrid, Spain, p. 228. Gabriels, D., 2000. Rain erosivity in Europe. In: Rubio, J.L., Ains, S., Andreu, V., de Paz, J.M., Gimeno, E. (Eds.), ESSC Third International Congress. Key Notes. Impression: Grupo Carduche, S. Cop. 46960, Aldaia, Valencia, Spain, pp. 31 43. Govers, G., 1987. Spatial and temporal variability in rill development processes at the Huldenberg experimental field. In: Bryan, R.K. (Ed.), Rill Erosion: Processes and Significance. Catena Supplement, vol. 8. Catena Verlag, Cremlingen, Germany, pp. 35 54. Govers, G., 1991. Rill erosion on arable land in central Belgium. Rates, controls and predictability. Catena 18, 133 155. Hasholt, B., Hansen, B.J., Olsen, C., Olsen, P., Sibbesen, E., 1997. Sediment delivery to streams from adjacent slopes on agricultural land in Denmark. In: Walling, D.E., Probst, J.L. (Eds.), Human Impact on Erosion and Sedimentation. IAHS Publ., vol. 245. IAHS Press, Centre for Ecology and Hydrology, Wallingford, UK, pp. 101 110. Hauge, C., 1977. Soil erosion definitions. Calif. Geol. 30, 202 203. Kamphorst, E.C., Jetten, V., Guerif, J., Pitkanen, J., Iversen, B.V., Douglas, J.T., Paz, A., 2000. Predicting depressional storage from soil surface roughness. Soil Sci. Soc. Am. J. 64, 1749 1758. Karssenberg, D., 1996. PCRaster Version 2 Manual. Department of Physical Geography, Faculty of Geographical Sciences, Utrecht University, Utrecht, The Netherlands. 368 pp. Kwaad, F.J.P.M., 1991. Summer and winter regimes of runoff generation and soil erosion on cultivated loess soils (The Netherlands). Earth Surf. Processes Landf. 16, 653 662. Ludwig, B., Boiffin, J., Chadoeuf, J., Auzet, A.V., 1995. Hydrological structure and erosion damage caused by concentrated flow in cultivated catchments. Catena 25, 227 252. Ludwig, B., Auzet, A.V., Boiffin, J., Papy, F., King, D., Chadoeuf, J., 1996. Etats de surface, structure hydro graphyque et erosion en rigole de bassins versants cultives du Nord de la France. Et. Gest. Sols 3 (1), 53 70. Nachtergaele, J., Poesen, J., 1999. Assessment of soil losses by ephemeral gully erosion using high-altitude (stereo) aerial photographs. Earth Surf. Processes Landf. 24, 693 706. Oygarden, L., 1996. Erosion and surface runoff in small agricultural catchments. In: Walling, D.E., Webb, B.W. (Eds.), Erosion and Sediment Yield: Global and Regional Perspectives. IAHS Publ., vol. 236. IAHS Press, Centre for Ecology and Hydrology, Wallingford, UK, pp. 251 266. Poesen, J.W.A., 1989. Conditions for gully formation in the Belgian loam belt and some ways to control them. In: Schwertmann, U., Rickson, R.J., Auerswald, K. (Eds.), Soil Erosion and Protection Measures in Europe. Soil Technology Series, vol. 1. Catena Verlag, Cremlingen-Destedt, Germany, pp. 39 52. Poesen, J.W.A., Govers, G., 1990. Gully erosion in the Loam belt of Belgium: tipology and control measures. In: Boardman, J., Foster, I.D.L., Dearing, J.A. (Eds.), Soil Erosion on Agricultural Land. Wiley, Chichester, UK, pp. 513 530. Poesen, J.W.A., Vandaele, K., Van Wesemael, B.L., 1996. Contribution of gully erosion to sediment production on cultivated lands and rangelands. In: Walling, D.E., Webb, B.W. (Eds.), Erosion and Sediment Yield: Global and Regional Perspectives. IAHS Publ., vol. 236. IAHS Press, Centre for Ecology and Hydrology, Wallingford, UK, pp. 251 266. ` Souchere, V., King, D., Daroussin, J., Papy, F., Capillon, A., 1998. Effects of tillage on runoff directions: consequences on runoff contributing area within agricultural catchments. J. Hydrol. 206, 256 267. Taboada, Ma.M., Paz, A., Valcarcel, M., 1999. Condiciones de formacion de la escorrenta en suelos de cultivo. n In: Paz-Gonzalez, A., Taboada Castro, Ma.T. (Eds.), Avances Sobre el Estudio de la Erosio Hdrica. Coleccion Cursos, Congresos e Simposios, vol. 52. Universidade da Coruna, A Coruna, Spain, pp. 69 99. Uhlen, B., 1986. Erosion of arable land in Sweden. Hordisk Hydrologisk Program Report 14, pp. 285 291. Valcarcel, M., 1999. Variabilidade espacial e temporal da erosion en solos de cultivo. PhD Thesis. EPS de Lugo, Univ. of Santiago de Compostela, Spain, p. 266. Vandaele, K., Poesen, J., 1995. Spatial and temporal patterns of soil erosion rates in an agricultural catchment, central Belgium. Catena 25, 213 226. Vansteelant, J.Y., Trevisan, D., Perron, L., Dorioz, J.M., Roybin, D., 1997. Conditions dapparition du ruisselle ment dans les cultures de la region lemanique. Relation avec le fonctionnement des exploitations agricoles. Agronomie 17, 65 82.