**1. Introduction**

150 Soil Erosion Studies

Toy, T., Foster, G., Renard, K., 2002. Soil erosion: Processes, Prediction, Measurement and

Vanesland, A., Lal, R., Gabreiels, D., 1987. The erodibility of some Nigerian soils: A

Vitharana, V., Mervenne, M., Simpson, D., 2008. Key soil and topographic properties to

Victora, C., Kaceevas, A., Flora, H., 1998. Soil erodibility assessments with simulated rainfall

Wei, W., Chen, L., Fu, B., 2007. The effect of land uses and rainfall regimes on runoff and soil erosion in the semi-arid loess hilly area, China, j. of Hydrology, 335, 247- 258. Yu, X., Zhang, X., Li, J., Zhang, M., Xie, Y., 2006. Effects of vegetation cover and precipitation

Zhang k., Li, S., Peng, W., Yu, B., 2004. Erodibility of agricultural soil on loess plateau of

Zhang, X., Liu, W., Li, Z., Zheng, F., 2009. Simulating site-specific impacts of climate change

Zhenge, F., 2006. Effect of Vegetation Changes on Soil Erosion on the Loess Plateau,

Zheng, M., Qiangguo, C., Hao, C., 2007. Effect of vegetation on runoff-sediment yield

Zhou, C., Shangguan, P., 2007. The effects of ryegrass roots and shoots on loess erosion

Zhou, C., Gan, T., Shangguan, P., Dong, B., 2010. Effects of grazing on soil physical

Congress of soil science. Montpellier. France. Symp. No. 31.

comparison of rainfall simulator result with estimates obtained from the wishmeier

delineate potential management class for precision agriculture in the European

and the USLE nomograph in soil from Uruguay. Proceedings of 16th World

on the process of sediment produced by erosion in a small watershed of loess

on soil erosion and surface hydrology in southern loess plateau of China, Catena,

relationship at different spatial scales in hilly areas of the loess plateau, North

properties and soil erodibility in semiarid grassland of the Northern Loess Plateau

Control, John Wiley & Sons Publication, 338p.

nomograph, Hydro. Process, 1, 255-265.

region, Acta Ecologica Sinica, 26(1), 1-8.

79, 237–242.

Pedosphere, 16(4), 420-427.

(China), Catena, 82, 87–91.

China, soil and tillage research, 76, 157-165.

China, Acta Ecologica Sinica, 27(9), 3572-3581.

under simulated rainfall, j. of Catena, 70, 350-355.

loess area, Geoderma, 143, 206-215.

#### **1.1 Plot-scale experimental studies: structure, equipment, hydrologic monitoring**

Plot-scale experimental studies are generally part of broader research projects aimed at improving the understanding of interrelations between processes involving hydrological, climatic and biological factors (Wainwright et al., 2000). Recently, these studies have become multidisciplinary, integrating fields such as hydrology, ecology and geomorphology. In a global environmental change and degradation context, plot-scale studies may provide information about runoff mechanisms, soil erosion and vegetation dynamics processes that result from these changes (Abrahams et al., 1995; Parsons et al., 1996). Furthermore, plotscale studies may focus on water fluxes and sediment transport processes at controlled conditions using rainfall simulation (Wainwright et al., 2000; Rickson, 2001). It is important to note that process control generally involves simplifying a complex system that is highly variable in time and space (Wainwright et al., 2000; Abrahams et al., 1998; Parsons et al., 1998). However, plot-scale studies have the advantage of allowing for detailed process monitoring at small scale, providing a basic description of the most relevant aspects (Michaelides et al., 2009).

Plot-scale studies are also useful in providing experimental data involving rainfall, surface runoff and soil erosion. These data are used as reference in modeling conception, calibration and validation. However, there can be considerable variability in soil erosion processes, as well as limitations of models atempting to simulate these complexities (Nearing, 2004). For example, in a study using 40 cultivated plots in the United States the experimental data coefficient of variation ranged between 18-91%. In addition, this variation was found to decrease with increasing rainfall erosive power (Wendt et al., 1986). Ruttimann et al. (1995) found that soil loss varied up to 173% between replicates under the same treatment. In general, the capacity of the model in representing local physical system can be tested by comparing observed and simulated model data, using regression analysis. Regression coefficient values from several studies demonstrate that model efficiency increases as erosion variability decreases, such as when mean annual soil loss data are used (Nearing, 1998; Risse et al., 1993; Zhang et al., 1996). The USLE-Universal Soil Loss Erosion (Wischmeyer & Smith, 1978) soil erosion model was originally conceived by using statistical

Plot-Scale Experimental Studies 153

Several plot-scale studies have also been developed under natural conditions. The aim of these studies is often to obtain data involving hydrologic and erosion processes and its relationship with biological factors, such as faunal and vegetation species dynamics (Reynolds et al., 1999). Although a reasonably long monitoring period is required, data obtained from these studies can produce a broader description of the actual system and

An experimental plot-scale study subjected to natural conditions was carried out in the semi-arid municipality of Serra Negra do Norte, northeastern Brazil (Moreira et al., 2009). The study aimed to analyze interrelationships between surface runoff, sediment transport and the aspects of undisturbed native vegetation in this region. The plot was installed at the Seridó Ecological Station (coordinates 6°34'42",S; 37°15'56",W), an environmentally protected area located at approximately 300 km from the city of Natal. Plot area was bounded by a 0.3 m height brick wall. Plot relief and situation within the Seridó Ecological Station catchment are shown in Figure 1. During the rainy season, around 70% of the plot area is composed of annual xerophyte species. Climate factors such as rainfall and temperature regulate water availability and biological processes in the region. In the drought season, plants are subjected to water stress. At the end of this period, the permanent species Mimosa tenviflora covers about 20% of the plot, along with substrate accumulated at the soil surface composed of leaves, seeds and dry twigs. In semi-arid areas, native vegetation plays an important role in infiltration and rainfall interception. For example, the stem structure direct flow to the plant roots, enhancing both soil infiltration and soil water storage. Similarly, residue and organic matter accumulated on the surface

existing interrelationships.

**2. Plot-scale experimental study in semi-arid Brazil** 

protect soil from the impact of raindrops (Puigdefábregas, 2005).

Fig. 1. Plot location within the catchment and terrain relief.

The average annual rainfall over the past 11 years (1995-2005) is 689 mm. Precipitation is highly variable from year to year in the area, where approximately 52% is high intensity thunderstorms of limited area extent. A study has shown the effect of high spatiotemporal

analysis of 20 years soil erosion data in natural and cultivated experimental plots, installed at 49 stations across the United States. It incorporated approximately 1000 events, producing a significantly representative database. In fact, USLE soil erosion model was developed primarily for agricultural purposes, in an atempt to simplify complex erosive interactions. Indeed, it paved the way for more refined modeling structures which consider the physical characteristics of the process.

Establishing an experimental plot often involves hydrological monitoring by using manual and automatic devices and fieldwork surveys to collect information on details such as plot topography, soil hydraulic characteristics, flora and fauna. In general terms, it is hypothesized that the plot represents local climate, soil and plant conditions. A plot-scale experimental study involving precipitation, surface runoff, soil erosion processes as well as the biological dynamics of local fauna and flora was developed in the semi-arid Brazilian Northeast (Moreira et al., 2009). Plot-scale studies often include topographic survey, analysis of soil surface characteristics such as roughness, crusting, cracking, and soil as an environment of biological activity for arthropods and other organisms. The plot is delimited and identifies the study area.

Hydrologic monitoring involves the measurement of variables, often requiring the installation of manual and automatic devices. Indeed, plot-scale studies in uncontrolled conditions tipically require the use of automatic devices. Water discharge monitoring implies the use of a measurement structure such as a Parshall flume or a tank at the downstream end of the plot. If a Parshall flume is used, discharge is monitored by using a stage-discharge relationship. Once the measuring structure is established, water surface monitoring is conducted by using manual (graduated rule) or automatic devices (water level logger). In case a tank is used, discharge is monitored by water surface variation as a function of time during the storm. After each event, the tank must be emptied and the sediment and particulate organic matter is collected, dried and analyzed. For each storm event, runoff was obtained by applying water balance equations including runoff, rainfall and the variation of the tank water level during the storm.

A plot-scale study was developed in New Mexico-USA using 15 small plots composed of 5 different grassland and shrubland species. The aim was to identify the hydrological and erosional processes resulting from these species in a context of degraded environment caused by the advance of shrub species in the region. 54 small-scale rainfall simulations (125 mm.h-1) found that shrub specie and canopy density were the main vegetation control on runoff and erosion. Significant interactions and feedbacks were found to occur between edaphic carachteristics and vegetation, which influenced both runoff and erosion responses (Michaelides et al., 2009; Wainwright et al., 2000).

Some researchers have highlighted the role of experimental studies at different scales, in light of the need to increase levels of complexity and connectivity in the study of processes (Bergkamp, 1998; Cammeraat, 2002). The results obtained on small-scale investigations present serious limitations and cannot be extrapolated to other scales without careful analysis. Experimental conditions at small-scale do not usually capture the interactions of a complex physical system (Kirkby, 1987; Zhang et al., 1999). Different processes can be dominant or observable at specific scales. For example, at a fine scale processes such as rain splash and rill and interill erosion are important, and at a larger scale, gully erosion, sediment deposition and other processes become more dominant (de Vente & Poesen, 2005).

analysis of 20 years soil erosion data in natural and cultivated experimental plots, installed at 49 stations across the United States. It incorporated approximately 1000 events, producing a significantly representative database. In fact, USLE soil erosion model was developed primarily for agricultural purposes, in an atempt to simplify complex erosive interactions. Indeed, it paved the way for more refined modeling structures which consider the physical

Establishing an experimental plot often involves hydrological monitoring by using manual and automatic devices and fieldwork surveys to collect information on details such as plot topography, soil hydraulic characteristics, flora and fauna. In general terms, it is hypothesized that the plot represents local climate, soil and plant conditions. A plot-scale experimental study involving precipitation, surface runoff, soil erosion processes as well as the biological dynamics of local fauna and flora was developed in the semi-arid Brazilian Northeast (Moreira et al., 2009). Plot-scale studies often include topographic survey, analysis of soil surface characteristics such as roughness, crusting, cracking, and soil as an environment of biological activity for arthropods and other organisms. The plot is delimited

Hydrologic monitoring involves the measurement of variables, often requiring the installation of manual and automatic devices. Indeed, plot-scale studies in uncontrolled conditions tipically require the use of automatic devices. Water discharge monitoring implies the use of a measurement structure such as a Parshall flume or a tank at the downstream end of the plot. If a Parshall flume is used, discharge is monitored by using a stage-discharge relationship. Once the measuring structure is established, water surface monitoring is conducted by using manual (graduated rule) or automatic devices (water level logger). In case a tank is used, discharge is monitored by water surface variation as a function of time during the storm. After each event, the tank must be emptied and the sediment and particulate organic matter is collected, dried and analyzed. For each storm event, runoff was obtained by applying water balance equations including runoff, rainfall

A plot-scale study was developed in New Mexico-USA using 15 small plots composed of 5 different grassland and shrubland species. The aim was to identify the hydrological and erosional processes resulting from these species in a context of degraded environment caused by the advance of shrub species in the region. 54 small-scale rainfall simulations (125 mm.h-1) found that shrub specie and canopy density were the main vegetation control on runoff and erosion. Significant interactions and feedbacks were found to occur between edaphic carachteristics and vegetation, which influenced both runoff and erosion responses

Some researchers have highlighted the role of experimental studies at different scales, in light of the need to increase levels of complexity and connectivity in the study of processes (Bergkamp, 1998; Cammeraat, 2002). The results obtained on small-scale investigations present serious limitations and cannot be extrapolated to other scales without careful analysis. Experimental conditions at small-scale do not usually capture the interactions of a complex physical system (Kirkby, 1987; Zhang et al., 1999). Different processes can be dominant or observable at specific scales. For example, at a fine scale processes such as rain splash and rill and interill erosion are important, and at a larger scale, gully erosion, sediment deposition and other processes become more dominant (de

characteristics of the process.

and identifies the study area.

and the variation of the tank water level during the storm.

(Michaelides et al., 2009; Wainwright et al., 2000).

Vente & Poesen, 2005).

Several plot-scale studies have also been developed under natural conditions. The aim of these studies is often to obtain data involving hydrologic and erosion processes and its relationship with biological factors, such as faunal and vegetation species dynamics (Reynolds et al., 1999). Although a reasonably long monitoring period is required, data obtained from these studies can produce a broader description of the actual system and existing interrelationships.
