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

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 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

Plot-Scale Experimental Studies 155

period began in the second half of January 2009 when events of substantial magnitude occurred. This period was followed by high-magnitude events in February and March 2009. During the rainy period, soil water moisture increased and annual plants showed progressive changes over a period of aproximately 2-3 weeks. As the rainy period continued, vegetation and biological activity interacted with soil, increasing porosity and enhancing soil storage capacity. Indeed, vegetation may act as sinks of overland flow and sediment due to the velocity reduction as runoff encounters plants (Ludwig et al., 2005). Collection campaigns showed a vigorous transformation in the natural landscape, with soil

Analysis of plot core samples provided quantification of the arthropod fauna and its effect on both seasonal periods. Organisms observed were classified into twelve different taxa: *Homoptera, Hemiptera, Hymenoptera, Coleoptera, Orthoptera, Psocoptera, Embioptera, Diptera, Collembola, Ácari, Araneae* and *Geophilomorpha*. Soil fauna demonstrated nine orders of insects, with a clear predominance of *Ácari* group, which occurred in the substrate mainly during the rainy season (Figure 3). It is important to note the mutual and positive relationship between the anual plant species and *Ácari, Collembola* and *Orthoptera*, found primarily in the organic substrate at the soil surface. The annual species offers suitable habitat for the arthropods, including provision of shade, refuge and food. However soil moisture is the most important factor for both the plants and arthopods. Accordingly, an increase in soil water moisture was followed by a marked increase in faunal activity during the rainy period in comparison with the drought period. The quantity of observed organisms in these 2 periods is classified by taxonomic rank and presented in Figure 3.

Fig. 3. Arthropod fauna observed in the plot during the dry and rainy seasons, classified by

*Embioptera*

*Geophilomorpha*

*Hemiptera*

*Homoptera*

*Hymenoptera*

*Orthoptera*

*Psocoptera*

With the aim of monitoring surface runoff and soil erosion processes, automatic devices (rainfall recorder and water level logger) were installed adjacent to the plot. A 5 m3 tank was

water availability as a crucial factor.

taxonomic rank.

**Number of Organisms**

*Acari (x10)*

*Araneae*

*Coleoptera*

*Collembola*

*Diptera*

Drought season Rainy season **Order**

**2.2 Hydrological monitoring and sediment yield**

rainfall heterogeneity on runoff in the watershed area (Moreira et al., 2006). Monthly rainfall shows considerable variation during the rainy season, especially in the period January-May. Daily rainfall data statistical analysis revealed that approximately 25% of the annual depth occurs during the maximum daily precipitation.

#### **2.1 Biological fauna in the soil and substrate**

The experimental plot-scale study enabled analysis of the biological fauna dynamics in the soil and substrate throughout the drought and rainy seasons in 2008-2009. For this purpose, five core sample collection campaigns were conducted at three randomly established points adjacent to the plot site. Soil samples were retrieved at 0.05 m depth and packed for subsequent laboratory analysis. Similarly, substrate samples were collected at three points, in squares measuring 0.30 x 0.30 m2. Material was then submitted for biological analysis at the UFRN Entomology Laboratory to identify the main arthropod groups observed in the soil and substrate. To that end, specific taxonomic identification keys were used [Zeppelini-Filho & Bellini (2004), Buzzi (2005), Triplehorn & Johnson (2005)]. Screening, counting and organism identification was conducted using a tray, metal tongs, Petri dish, test tube and sterile microscope. Organisms not identified in the previous phase were removed with a Berlese-Tullgren funnel. After counting, arthropods were then stored in a test tube containing alcohol at 70° (v/v). Figure 2 presents daily precipitation data on the plot during the study period of 2008-2009. Samples were collected during drought and rainy periods.

Fig. 2. Daily precipitation and sample collection as a function of time.

It is observed that after a period of several precipitation events during the first half of April 2008, further precipitation events occurred for the next couple of months. The first collection campaign was conducted on 08/08/2008. This was preceded by a period of sporadic lowmagnitude rainfall, which produced a deficit on soil moisture and significant effects on annual species. The period between June and December 2008 received close to zero precipitation, with only an isolated event (13 mm) on 12/12/2008. A well-defined 165-day drought was observed during this period, which increased soil water stress level. The rainy

rainfall heterogeneity on runoff in the watershed area (Moreira et al., 2006). Monthly rainfall shows considerable variation during the rainy season, especially in the period January-May. Daily rainfall data statistical analysis revealed that approximately 25% of the annual depth

The experimental plot-scale study enabled analysis of the biological fauna dynamics in the soil and substrate throughout the drought and rainy seasons in 2008-2009. For this purpose, five core sample collection campaigns were conducted at three randomly established points adjacent to the plot site. Soil samples were retrieved at 0.05 m depth and packed for subsequent laboratory analysis. Similarly, substrate samples were collected at three points, in squares measuring 0.30 x 0.30 m2. Material was then submitted for biological analysis at the UFRN Entomology Laboratory to identify the main arthropod groups observed in the soil and substrate. To that end, specific taxonomic identification keys were used [Zeppelini-Filho & Bellini (2004), Buzzi (2005), Triplehorn & Johnson (2005)]. Screening, counting and organism identification was conducted using a tray, metal tongs, Petri dish, test tube and sterile microscope. Organisms not identified in the previous phase were removed with a Berlese-Tullgren funnel. After counting, arthropods were then stored in a test tube containing alcohol at 70° (v/v). Figure 2 presents daily precipitation data on the plot during the study period of 2008-2009. Samples were collected during drought and rainy periods.

occurs during the maximum daily precipitation.

**2.1 Biological fauna in the soil and substrate** 

Fig. 2. Daily precipitation and sample collection as a function of time.

0

20

40

**P (mm)** 

60

80

100

120

It is observed that after a period of several precipitation events during the first half of April 2008, further precipitation events occurred for the next couple of months. The first collection campaign was conducted on 08/08/2008. This was preceded by a period of sporadic lowmagnitude rainfall, which produced a deficit on soil moisture and significant effects on annual species. The period between June and December 2008 received close to zero precipitation, with only an isolated event (13 mm) on 12/12/2008. A well-defined 165-day drought was observed during this period, which increased soil water stress level. The rainy

**Time** Sample collection

period began in the second half of January 2009 when events of substantial magnitude occurred. This period was followed by high-magnitude events in February and March 2009. During the rainy period, soil water moisture increased and annual plants showed progressive changes over a period of aproximately 2-3 weeks. As the rainy period continued, vegetation and biological activity interacted with soil, increasing porosity and enhancing soil storage capacity. Indeed, vegetation may act as sinks of overland flow and sediment due to the velocity reduction as runoff encounters plants (Ludwig et al., 2005). Collection campaigns showed a vigorous transformation in the natural landscape, with soil water availability as a crucial factor.

Analysis of plot core samples provided quantification of the arthropod fauna and its effect on both seasonal periods. Organisms observed were classified into twelve different taxa: *Homoptera, Hemiptera, Hymenoptera, Coleoptera, Orthoptera, Psocoptera, Embioptera, Diptera, Collembola, Ácari, Araneae* and *Geophilomorpha*. Soil fauna demonstrated nine orders of insects, with a clear predominance of *Ácari* group, which occurred in the substrate mainly during the rainy season (Figure 3). It is important to note the mutual and positive relationship between the anual plant species and *Ácari, Collembola* and *Orthoptera*, found primarily in the organic substrate at the soil surface. The annual species offers suitable habitat for the arthropods, including provision of shade, refuge and food. However soil moisture is the most important factor for both the plants and arthopods. Accordingly, an increase in soil water moisture was followed by a marked increase in faunal activity during the rainy period in comparison with the drought period. The quantity of observed organisms in these 2 periods is classified by taxonomic rank and presented in Figure 3.

Fig. 3. Arthropod fauna observed in the plot during the dry and rainy seasons, classified by taxonomic rank.

#### **2.2 Hydrological monitoring and sediment yield**

With the aim of monitoring surface runoff and soil erosion processes, automatic devices (rainfall recorder and water level logger) were installed adjacent to the plot. A 5 m3 tank was

Plot-Scale Experimental Studies 157

<sup>0</sup> ( ) ( ). *k t c c f t f ff e*

where f(t) represents infiltration capacity at time t (L.T-1), fo is the initial infiltration rate (L.T-1), fc is a final infiltration capacity (L.T-1) and k is an empirical constant. Horton equation parameters reflect the spatial heterogeneity of soil hydraulics. In addition, the average observed saturated hydraulic conductivity rates in the rainy period are approximately six times higher than in the drought period, which indicates a marked difference in soil hydraulic behavior. Indeed, during the rainy season soil infiltration capacity is enhanced by an increase in soil moisture, roots osmotic effect, vegetation cover and faunal activity. Higher soil infiltration capacity rates were observed in areas beneath the canopy of permanent species such as *Mimosa tenviflora* (medium-sized trees). In these areas, a higher density of annual plants was observed, mainly due to canopy shade which

provides protection from high temperatures and radiation.

1 30 4

Table 2. Infiltration parameters of the Horton equation (mm.h-1).

Experimental run

Drought period

**2.4 Runoff generation mechanisms** 

periods in Figures 4(a) and 4(b), respectively.

.

Parameters Experimental

2 25 3 2 160 85 3 120 50 3 180 120 4 60 6 4 35 10 5 35 10 5 240 170 6 36 12 6 170 155 7 30 5 7 180 160 8 80 57 8 90 40

Rainy period

Surface runoff and erosion in semi-arid areas are the result of various factors associated with rainfall (duration and intensity), soil (moisture, cracking, crusting, and soil infiltration capacity), plant cover (density) and terrain relief. During the study period, 46 precipitationrunoff-sediment yield events were recorded and their main characteristics and the corresponding hydraulic responses were evaluated. 55% of the rainfall events duration were less than 60 minutes in duration and 66% between 18h00 and 06h00. Rainfall peak rate ranged between 9.14 and 137.16 mm.h-1. In 56% of events peak rate surpassed 40 mm.h-1 and in 5 events it exceeded 90 mm.h-1. The runoff coefficient is the relationship between surface runoff and rainfall levels during the event. Observed values of runoff coefficient, rainfall peak rate and precipitation height are presented for the beginning and the end of the rainy

33% of the observed events didn't produce runoff. Observed runoff coefficients were lower than 0.1 for 82% of events, which indicate high soil water storage capacity. Only 5 events exhibited runoff coefficients higher than 0.2; these higher values were possibly due to influence of antecedent rainfall, soil water storage capacity and the density of vegetation cover.

(2)

run

fo fc fo fc

Parameters

1 300 195

built at the downstream end of the plot in order to collect water discharge and sediment flowing from the plot during each event. The water level logger was programmed to take measurements every 5 minutes. After each event, the tank was emptied using a portable pump and sediment deposited at the bottom was collected, dried, weighed and analyzed. Calculation of surface runoff in the plot included precipitation and the variation of the tank water level during the storm, in accordance with the water balance Equation 1,

$$RO = \frac{1}{\Delta t} \left[ A\_{sw} (h\_t - h\_{t-1}) - (P\_{sw}, A\_{sw}) - (P\_{ramp}, A\_{ramp}) \right] \tag{1}$$

where RO (L3.T-1) is the surface runoff; Δt (T) is the interval between measurements; Psw and Pramp are the precipitation height on tank water surface and the paved ramp (L), respectively; Asw e Aramp are the tank and ramp water surface areas (L2), respectively; ht and ht-1 are the surface water levels in the tank (L) at times t and t-1, respectively.

**2.3 Soil hydraulic properties** 

Soils in the plot site can be classified as a variation between clay-gravel textured Chromic Luvisol and the expansive clay Vertisoil. The soils in the plot site are well drained, shallow, gravelly, with depth varying from 0 to 1.40 m, with rocky outcrops on approximately 15% of its surface. Soil samples were collected from the top 0.1 m and taken to the laboratory for size distribution analysis, which allowed determination of the representative diameters D50, D16, D84 (in milimeters) and standard deviation, whose values are presented in Table 1.


Table 1. Soil representative diameters from the plot.

Table 1 indicates that soil size distribution is composed of gravel, fine, medium and coarse sand, silt and clay. Standard deviation values indicate bimodal composition of the soil (fine and coarse modes present in the bulk sample). The natural roughness of the surface area during the dry season is mainly due to the occurrence of randomly crusted rock fragments, which results from both physical and chemical weathering processes. During the first rainstorm events, readily mobilized sediment dominate sediment yield. Seeds and organic matter are also transported by overland flow across the plot.

To determine soil hydraulic properties in the plot, field infiltration experiments were conducted during the drought and rainy periods. Experiments were performed using a constant head disc permeameter at sixteen points, with care taken to avoid disturbing the native vegetation. The objective was to investigate vertical flow behavior through soil profile as a function of time. The field infiltration experimental data was used to adjust the Horton infiltration parameters, which characterise soil hydraulic properties as an unsaturated porous media. Infiltration curves exhibit declining behavior, with a constant and asymptotic tendency as a function of time. Thus, the profile achieved steady regime at soil saturation level. The Horton infiltration equation (1933) is as follows,

built at the downstream end of the plot in order to collect water discharge and sediment flowing from the plot during each event. The water level logger was programmed to take measurements every 5 minutes. After each event, the tank was emptied using a portable pump and sediment deposited at the bottom was collected, dried, weighed and analyzed. Calculation of surface runoff in the plot included precipitation and the variation of the tank

water level during the storm, in accordance with the water balance Equation 1,

1

and ht-1 are the surface water levels in the tank (L) at times t and t-1, respectively.

**2.3 Soil hydraulic properties** 

Table 1. Soil representative diameters from the plot.

matter are also transported by overland flow across the plot.

soil saturation level. The Horton infiltration equation (1933) is as follows,

where RO (L3.T-1) is the surface runoff; Δt (T) is the interval between measurements; Psw and Pramp are the precipitation height on tank water surface and the paved ramp (L), respectively; Asw e Aramp are the tank and ramp water surface areas (L2), respectively; ht

Soils in the plot site can be classified as a variation between clay-gravel textured Chromic Luvisol and the expansive clay Vertisoil. The soils in the plot site are well drained, shallow, gravelly, with depth varying from 0 to 1.40 m, with rocky outcrops on approximately 15% of its surface. Soil samples were collected from the top 0.1 m and taken to the laboratory for size distribution analysis, which allowed determination of the representative diameters D50, D16, D84 (in milimeters) and standard deviation, whose values are presented in Table 1.

Sample A1 A2 A3 A4 A5 A6 A7 D84 1.8 0.6 0.43 0.55 0.6 0.39 0.49 D16 0.04 0.029 0.035 0.053 0.053 0.049 0.035 D50 0.17 0.06 0.15 0.17 0.17 0.15 0.15 S.D. 6.71 4.55 3.51 3.22 3.36 2.82 3.74

Table 1 indicates that soil size distribution is composed of gravel, fine, medium and coarse sand, silt and clay. Standard deviation values indicate bimodal composition of the soil (fine and coarse modes present in the bulk sample). The natural roughness of the surface area during the dry season is mainly due to the occurrence of randomly crusted rock fragments, which results from both physical and chemical weathering processes. During the first rainstorm events, readily mobilized sediment dominate sediment yield. Seeds and organic

To determine soil hydraulic properties in the plot, field infiltration experiments were conducted during the drought and rainy periods. Experiments were performed using a constant head disc permeameter at sixteen points, with care taken to avoid disturbing the native vegetation. The objective was to investigate vertical flow behavior through soil profile as a function of time. The field infiltration experimental data was used to adjust the Horton infiltration parameters, which characterise soil hydraulic properties as an unsaturated porous media. Infiltration curves exhibit declining behavior, with a constant and asymptotic tendency as a function of time. Thus, the profile achieved steady regime at

<sup>1</sup> ( )( . )( . ) *RO A h h P A P A sw t t sw sw ramp ramp <sup>t</sup>* (1)

$$f(t) = f\_c + (f\_0 - f\_c).e^{-k.t} \tag{2}$$

where f(t) represents infiltration capacity at time t (L.T-1), fo is the initial infiltration rate (L.T-1), fc is a final infiltration capacity (L.T-1) and k is an empirical constant. Horton equation parameters reflect the spatial heterogeneity of soil hydraulics. In addition, the average observed saturated hydraulic conductivity rates in the rainy period are approximately six times higher than in the drought period, which indicates a marked difference in soil hydraulic behavior. Indeed, during the rainy season soil infiltration capacity is enhanced by an increase in soil moisture, roots osmotic effect, vegetation cover and faunal activity. Higher soil infiltration capacity rates were observed in areas beneath the canopy of permanent species such as *Mimosa tenviflora* (medium-sized trees). In these areas, a higher density of annual plants was observed, mainly due to canopy shade which provides protection from high temperatures and radiation.


Table 2. Infiltration parameters of the Horton equation (mm.h-1).
