Soil Erosion and Conservation Measures in Semiarid Ecosystems Affected by Wildfires

Fire is a factor of the first order in the interpretation of current ecosystem composition and landscape shape. Wildfires are a common phenomena dating back to the Earth’s origins when they were caused by volcanic events or lightning. However, nowadays fires are more often caused by human interventions, accidental or deliberate, than by natural ones. In the European Mediterranean regions wildfires increases from 1960s aided by a general warming and drying trend, but driven primarily by socio-economic changes, including rural depopulation, land abandonment and afforestation with flammable species (Shakesby, 2011). Within that global Mediterranean region, semiarid areas have plant communities of sclerophyllous evergreen shrubs, rich in essences and with low plant water content, especially in dry seasons. These conditions facilitates the burning of these ecosystems. Wildfires can be considered as a disturbance with a relatively severe temporary impact (Cerdà & Doerr, 2005). Their severity will depend on key factors such as the intensity and frequency of fire, but also on the the ecosystem components (vegetation, soil, rainfall, etc). The burning of plant cover and litter leaves the soil unprotected against rain impact. Also, immediately following a wildfire, a layer of ash and charred material typically covers the ground. It can increase or decrease the post-fire runoff and erosion response, depending upon the soil and ash properties and the ash thickness (Bodí et al., 2011). Soil loss after wildfire increases as long as there is no a minimal recovery of both plant cover and soil properties (Mataix et al., 2007). Although some strategies, such as plant resprouting, allow them to recover quickly, high fire frequency or soils with low quality can requier restoration strategies (Cerda & Robichaud, 2009); these aspects that will be discused in this chapter.

these effects by means of two field methods: (i) sediment traps or Gerlach boxes and (ii) a rainfall simulator. Sediment traps give us a continuous measurement of soil erosion in microplots. A total of six Gerlach collectors (50 x 16 x 16cm) were installed under different vegetation covers and limited in area of influence with foil of a meter in length. Rainfall simulators give us a measurement of sheet erosion in burned microplots against a rain of given intensity (80 mm h -1 ) in comparison to paired unburned areas (n=9). Experiments were carried out after fire for different soil units, such as Calcaric Regosol developed on Oligocene marls (calcareous soils) and Rendzic Phaeozem developed on colluvium (colluvial soils) as well as gypsiferous soils, classified as Haplic Gypsisol.

Results and discussion
Runoff and soil loss increased significantly after fire with exposure to strong rainfall (80 mm h -1 ) in a Calcaric Regosol below a Pinus halepensis forest (Fig.1). The increase in soil erosion and runoff after fire is related to the decrease of soil surface cover, as well as in the soil aggregate stability on the surface by heat effect . Also, the different soil characteristics generate different responses closely related to their parent material, thus the erosion of soils developed on marls is greater than those developed in gypsum and colluvium (Fig. 2). This relates to the low infiltration of marly soils, as well as temporary surface alteration due to fire (Cerdà & Doerr, 2005;Giovannini et al., 1990). Gypsiferous soils, with a lower fuel availability to alter the soil surface, and calcareous colluvial soils with more organic matter, stony and well-structured, did not reveal many problems of sheet erosion. The influence of the parent material of burned soils is shown in Figure 2 where we see a different curve for each soil type tested and confirm that the differences in soil characteristics generate different responses. The variability of these and other results is due to the diversity of variables under field conditions and confirms that the differences in soil characteristics generate different responses (Badía & Martí, 2003a). A compilation of tests performed with rainfall simulators in similar semi-arid Mediterranean environments can be found in Table 1. www.intechopen.com In burned soils with a bare surface and therefore with the soil surface altered, the runoff coefficient (~90% of applied water) and erosion (~850 g m -2 h -1 ) are very high. In different localites, the soil loss was lower in unburned plots (~0,2 Mg ha -1 year -1 ) than in burned plots, which accounted for between 2 and 20 Mg ha -1 year -1 over the first years. The following relationship between erosion and plant cover was established after the fire.
Sheet erosion (g m -2 ) = 282 -140,1 (Plant cover, %); r = 0,97, n = 10 (1) The relevance of plant composition on plant cover evolution after fire and the effects on soil erosion has been demonstrated (Fig. 3). So, soil loss was significanty lower in a dense Quercus coccifera L. shrubland than in an open Pinus halepensis L. forest in Castejón de Valdejasa. The figure 3 shows as plant community structure, as well as the sprouters:seeder ratio, explains differences in bare soil surface (Y2) and their relationship with soil loss (Y1) during the first months (X, time) after a wildfire.  ); Tr, Time to runoff (s); Sc, Sediment concentration (g L -1 ); Wf, Wetting front (cm); EC, Electrical conductivity of runoff (dS m -1 ) of overland flow. Table 1bis. Runoff and erosion data with rainfall simulator method in semi-arid Mediterranean ecosystems (arranged by publication date).
In Figure 3, there are also some turning points in the vegetation cover and erosion evolution related to environmental conditions. Thus, in spring vegetation the cover rate is accelerated then decreases during the summer period mainly dominated by therophytes plants; the pulses in the soil erosion are related to the greatest intensity of rainfall. It must be taken into account that in burned kermes evergreen oak shrubland, the average erosion is 7 times higher than in unburned plots; in Aleppo pinewood soil loss can be as much as 36 times higher in the early years . This is due to the differential evolution of vegetation to cover the soil surface after the fire.

Rainfall energy effects on soil erosion and runoff generation in semi-arid forested lands
The effects of fire and torrential rainfall, two important factors in Mediterranean semiarid environments, on the soil erosion and hydrology, are analyzed in this section.

Methods
Experiments were carried out using a portable sprinkler-based rainfall simulator with different nozzles and pressures. Two levels of rainfall energy (12,6 J m -2 mm -1 and 24,7 J m -2 mm -1 ) and similar intensity (85±8 mm h -1 ) reproduce torrential rainfall characteristics. Rainfall simulations were conducted immediately after Aleppo pine (Pinus halepensis L.) litter covering soil surface was burned. The experiments were conducted on calcareous, loamy clay soils classified as Calcaric Regosol, frequent in the semi-arid Central Ebro Valley (NE-Spain). Paired burned areas were established in nine micro-plots and compared with paired not burned control areas (2-soil status x 2 rainfall energy x 9 plots or replicates). In each rainfall simulation the soil loss, soil infiltration (calculated by Horton model), wetting front, runoff coefficient and runoff quality (EC and pH) were measured .

Results and discussion
The results indicate that when litter cover was burned rainfall significantly increases the sediment yield. Burned plots generated 18.5 times more sediment than unburned plots (control) for fine rainfall energy (Raindrop D 50 = 1,0 mm) and 33.6 times more for coarse rainfall energy (Raindrop D 50 = 1,4 mm) (Fig. 4). The return period for a single storm with I 30 (rainfall intensity for 30 min) ranged from 5 to 200 years from the East Mediterranean areas to semiarid Central Ebro Valley, respectively (Fomento, 2001). These rainfalls of high intensity are especially frequent in autumn, just after summer, the season of maximum risk of wildfires. Sediment loss was given as solutes dissolved in overland flow, mainly in unburned plots, and as particles in suspension, mainly in burned plots. However, soil erosion rates were without significant differences between both rainfalls. Fire increased runoff quantity (about 1,6 times) and decreased its quality temporarily by increasing significantly both pH and specially EC in relation to unburned plots (Fig. 5).
The results indicate that when vegetation and litter cover of soils were burned, the first rainfalls duplicate runoff, soil infiltration decreases significally and soil erosion increases 20 -30 times in relation to unburned plots, especially with the highest rainfall energy. Because of the dynamic climate of the area that we discussed, a restoration strategy for a short-term response would be an interesting response, especially in conditions where plant succession is very slow, e.g., high slopes and soils of low permeability materials, thus highly erodible. www.intechopen.com

Fire effects on soil properties under laboratory-controlled heatings
Wildfire passage is accompanied by a heat wave and a deposition of ash as a result of biomass combustion (Fig. 6). Fire effects are directly related to the soil type affected as well as fire severity, i.e. the heat reached and the quantity and quality of ashes deposited on soil surface. To test these variables we treated two contrasting soils under laboratory conditions, the most frequently found soils in the semi-arid Central Ebro Valley (NE-Spain), a Calcaric Regosol developed on marls and a Haplic Gipsisol, developed on gypsiferous marls.

Methods
Samples of both soils (calcareous and gypsiferous soil) were collected randomly and analyzed separately with 4 replicates for each soil type. Soil samples were heated for 30 minutes in a muffle furnace at temperatures of 25°, 150°, 250° and 500°C. These temperatures cover the range that is habitually reached in surface soils affected by fires (Chandler et al., 1983;Giovannini & Lucchesi, 1997;Walker et al., 1986). Soil sample treated at 250ºC was selected for adding black ashes, in a quantity related to plant biomass growing on each soil (Martí, 1998). The amount of ashes added to calcareous soil was twice (10 g kg -1 ) that added to gypsiferous soil (5 g kg -1 ).

Results and discussion
Similar changes in chemical properties were found for both soils at highest temperatures but differences between soils were observed at intermediate temperatures ( Among the results it should be noted that heating soil to 250°C caused a decrease in pH and an increase in electrolytic conductivity (ECe) and soluble Ca. Heating soil to 500°C caused an increase in pH and a decrease in ECe and soluble Ca. Increasing heat intensity increased organic matter loss by combustion, which was accompanied by an increase of available nutrients content. Total N content decreased at temperatures greater than 250°C, with about one-third being volatilized. Cation exchange capacity (CEC) was reduced for gypsiferous soil heated to 500°C and to 250°C for calcareous soil. As for physical soil properties, heating increased the quantity of sand-sized particles by fusion of clay with www.intechopen.com the greatest increase occurring in soil heated to 500°C. Soil aggregate stability (SAS) of both soils was reduced by heating to 250°C with greater reductions at 500°C, likely due to a reduction in organic matter and clay size particle content. Bulk density and particle density increased in both soils when heated to 500°C. Water availability (difference between field capacity and permanent wilting point) increased when soils were heated to the highest temperature, likely due to texture and structural modifications. The addition of ashes increased organic matter content, C/N ratio, and pH in both soils and increased nutrient availability, especially in the calcareous soil where soil addition was higher than in gypsiferous one. Soil respiration were quickly enhanced in calcareous soil but depleted in gypsiferous soil for intermediate heating treatments (150°C and 250°C). At the highest temperature (500°C), these biological properties were significantly reduced in both soil types and on a long term basis. Black ash addition increased basal respiration in both soils but did not affect other biological properties. These results demonstrate the existence of both labile and permanent effects of soil burning and a differential response to C dynamics as a function of soil properties (Badía & Martí, 2003b). The unburned or control soils (25ºC) have originally contrasting properties (C versus G soil) that are not strongly altered at intermediate temperatures (150ºC, 250ºC). But the highest heat treatment (500ºC) conducted on both soils (C500 vs G500) showed a lot of similarities as result of the degradation of their initial properties (Fig. 7). Fig. 7. Dendrogram of experimental calcareous (C) and gypsiferous (G) soils at different temperatures (25ºC, 150ºC, 250ºC, 500ºC) and ash addition (A). Numbers 0 to 25 shows the rescaled distance cluster combine, and 1 to 10 are the case numbers.

Methods
Water repellency is a soil property that affects its hydrological response and its erodibility (Cerdà & Doerr, 2008;Jordán et al., 2010). In this chapter, the effect of an experimental fire is studied on the soil water repellency of three different ecosystems. Three experimental soils have been selected in the NE of Spain with different properties (Table 3).  Table 3. Some characteristics of the studied soils and their ecosystems.
Non-altered blocks of soil have been obtained in the field in every area. All treatments have been carried out three times. Once in the laboratory, all the soil blocks have been air-dried and they have been burnt with a blowlamp (Llovet et al., 2008) up to reaching a maximum of 250ºC at 1 cm depth (Fig. 8). Fig. 8. Soil blocks were burned with a blowlamp reaching 250ºC at 1 cm soil depth.

www.intechopen.com
In order to evaluate the soil water repellency, the water drop penetration time (WDPT) test has been used, applying three drops per soil block. In addition, the water repellency variation is assessed in three different levels of depth (at the surface, at 2 cm and 5 cmdepth). The water repellency classification criterion is the one used by Doerr et al., 2009).

Results and discussion
Fire affect soil water repellency in a different way according to soil type and soil depth (Table 4). Fire decreased significantly the presence of water repellency on the surface of all experimental soils: Phaeozems, Calcisols and Cambisols. The fire reduced water repellency even in depth in Phaeozems (to 5 cm depth) and Calcisols (to 2 cm depth). Unlike, fire increased water repellency in Cambisols at 5 cm depth, from hydrophilic (unburned soil samples) to strongly hydrophobic (burned soil samples).  Waxes and similar lipid materials could be mostly responsible for the original soil water repellency, which can be enhanced or reduced according to soil properties (Fig. 9). The increase in water-repellent conditions at 5 cm depth in Cambisol could be attributed to the www.intechopen.com translocation into the soil depth of lipid fractions released from burning organic matter or biomass (González et al., 2004).

Rehabilitation practices in areas affected by wildfires
Some rehabilitation practices have been implemented as urgent measures to control soil erosion in areas affected by wildfire in thesemiarid Central Ebro Valley ( Figure 10): rehabilitation practices include herb seeding, re-mycorrhization and mulching with barley straw and pine woodchip. The effectiveness of these treatments are discussed in this section.

Methods
In a paired plots design, treatments of seeding, seeding and mycorrhized, seeding and mulching, and control (untreated) erosion plots were established in four different slopes and on calcareous one (Calcaric Regosols) and gypsiferous (Haplic Gipsisols) soils (Fig. 11). All sites were north-facing with a slope of around 20º. The seeding treatment consisted of the addiction of legumes and grass seed mixture, either native or established through cultivation in the region: Medicago sativa L., Medicago truncatula Gaertn., Onobrychis viciifolia Scop., Vicia villosa Roth, Agropyron cristatum (L.) Gaertn., Dactylis glomerata L., Lolium rigidum Gaud. and Phalaris canariensis (L.). Planting was done manually in rows, perpendicularly to the maximal slope (parallel contour seeding). The sowing rate for the whole mixture was 30 g m -2 , with an equivalent weight for each species. Mulching consisted of barley (Hordeum vulgare L.) straw covering the soil surface at a 100 g m -2 rate (Badía & Martí, 2000). And mycorrhization has consisted of contribution of vesicular-arbuscular mycorrhizae (Glomus mossae) to increase the nutrient and water potential of sown plants (Mataix et al., 2007;Vallejo, 1996).

Results and discussion
Plant cover (native and sown plants) in both soils (calcareaous and gypsiferous) after four months of seeding treatments are shown (Fig. 12). Native plant coverage in calcareous soils (40%) is greater than in gypsiferous soils (10%) where physical, chemical and biological properties are relatively worse. Moreover, in calcareous soils the resprouters are higher than in gypsiferous one. In gypsiferous plots the principal native species was Helianthemum syriacum while Brachypodium retusum was dominant in calcareous plots, in some cases reaching 50% of total plant cover. It is important to remark that plant cover of seeded plants is similar in both soils (~30%). Among the introduced species Lolium rigidum covered about 12% of soil surface. The other seeded species had a similar importance, 5% of the total for Onobrychis vicifolia and about 2-4% for each of the other six species.
Regarding the effectiveness of applied measures on soil erosion, we found differences between soils because bare soil after treatments was about 45% in gypsiferous soils and only 15% in calcareous one (Fig. 9). The soil loss in gypsiferous soils twice that one in calcareous soils (Table 4). The cumulative annual erosion rate measured in Gerlach boxes was about 4 Mg ha -1 year -1 for gypsiferous soils, 3 times more than treated soil. And in calcareous soils, soil losses was about 2 Mg ha -1 year in control plots were about 2 times higher than treated plots (  Table 5. Average erosion rate (n=4) in Gerlach collectors under different soil treatments: burned control plots (C), seeded plots (S), seeded and mycorrhized plots (SM), and seeded and mulched with straw (SMS). The plots were monitorized during two years after wildfire in Fraga hillslopes (NE-Spain). Different letters between treatments and soils indicates significant differences (LSD test, P<0,05).
The parallel contour seeeding treatment increases soil cover and acts as a sucessive microbarriers to the maximal slope to reduce flow velocity and runnoff coefficient while increasing infiltration and trapping sediments. This was verified in experiments with rainfall simulator . The runoff coefficient was significantly lower for seeding plots (13 to 38%) than for control soils (45 and 58% for calcareous and gypsiferous soils respectively). The lower calcareous soils runoff is associated with better physical and chemical properties than gypsiferous soils, differences in soil characteristics which are also reflected in the vegetation cover (Fig. 9). Because of the heterogeneity of rainfall in the region, the erosion is generated in the most rainy months of the study on both the gypsiferous and calcareous soils, and then is when there are significant differences between seeding and control plots. The Figure 13 note these differences on gypsiferous burned soils, and is representative also for calcareous soils. Fig. 13. Sediment yield (g m -2 ) under different treatments (control and seeding) on gypsiferous burned soils (n=4) and monthly rainfall (mm) along the period of study. Asterisks indicate the significant differences between seeding and control plots in the reporting month. From Badía & Martí, 2000. Cumulative soil loss was higher in gypsiferous soils than in calcareous soils for control plots (untreated). Although the application of any of the treatments reduced these rates and the www.intechopen.com benefits are clearly apparent, the erosion levels of post-fire degradation with no treatment cannot be classified as severe. These facts are consistent with other studies where wildfires are considered as a perturbation with a relatively severe temporary impact (Cerdá & Doerr, 2005). Analyzing the vegetal cover from the point of view of the pastoral value , fresh and dry plant matter increased with sowing as well as non-nitrogenous products. Straw mulching is the most effective treatment against soil erosion, inducing some improvement both in quantity and quality of plant matter in both types of soils. Remycorrhization did not affect these parameters. Endomycorrhizae recovery after the fire has been rapid (Table 4), hence the low response of a mycorrhizal inoculum allochthonous Martí, 1994a and1994b). Although at intermediate temperatures soil microbial activity is quickly enhanced, for specific organisms, such as endomycorrhizal fungi, linked to the recovery of vegetation, this recovery is slower (Table 6).  Table 6. Characterization of endomycorrhizal state (x  sd) in burned soils (n=4) one year after wildfire in NE-Spain (Barceló et al., 1994).

Species Survival and/or growth Location /Rainfall Years Author
P. halepensis S: 94,5% (90,7-97,5). HGR=35,8cm/year (30,3 (Martí & Badía, 1995). The low survival in calcareous soils (more fertile than gypsiferous soils), is due to competition with other species, especially with Brachypodium ramosum with which it has been important. In fact, the mortality of Pinus halepensis is inversely related to vegetal recovery, which highlights the competition for soil water reserves: Mortality, % = 2,054 (Vegetation cover, %) -23,513 r 2 = 0,578; P<0,01 More evidence of the role of soil water as a limiting factor in the study area, is that when reforestation practices are accompanied by a support irrigation in the plantation phase of the species, the survival rate and growth rises significantly. This was observed in a study of reforestation of opencast surfaces in the Val of Ariño coalfield (Teruel, NE-Spain) (Badía et al., 2007). In this study, the effect of a slope gradient on the growth and survival of different species 10 years after the reconstruction of mining banks was evaluated. Their survival shows significant differences: 95% Pinus halepensis, 81% Pistacia lentiscus, 62% Juniperus phoenicea, 41% Quercus ilex and 34%Quercus coccifera. Aleppo pine has been the fastest growing both height (36 cm year -1 ) and basal diameter (10 mm year -1 ), because of its greater adaptation in semiarid environments. For the other species, height values are less than 12 cm year -1 and 2,1 mm year -1 for basal diameter values. Growth and survival obtained here were higher than those of the same species in other afforestations in semiarid conditions. This outcome demonstrates the adequacy of species and applied techniques of restoration, in particular irrigation of summer season support, that allow a long-term reliability of reclaimed mine slopes.

Conclusions
In the semi-arid environments of NE-Spain, there are different types of soil conditioned by the lithology of the substrate derived, and whose erosion reponse is significantly divergent. Thus the calcareous burned soils with low stoniness and loamy clay texture, as well as eroded gypsiferous soils, show runoff and erosion levels higher than gypsiferous soils, and especially that colluvial soils. The coverage provided by vegetation is a key element to stop sheet erosion on these soils. The combined effect of heat and ash incorporation involves a number of changes in physical, chemical and biological characteristics of soils affected by wildfire. These effects vary according to intensity of the fire, the amount of ash incorporated and soil characteristics starting. The protective measures such seeding and mulching treatments involve a significant reduction of sediment yield in the first years after fire, higher in eroded gypsiferous soils than in fertile calcareous ones. The remycorrhization is not effective when the soil mycorrhizal status has already been recovered from the effects of fire.
In the forest domain of the area there is a diversity of species whose response after the fire is clearly different. Those with resprouting shrub species tend to recover its previous status faster, the vegetal recovery is higher and temporarily stable. In mature Aleppo's pine forest, the germination of pine is high; the pine density is 3-times higher before than after fire, although it will be prgressively reduced. The shrub and woody species introduced show a differential response according to the soil in which they are implanted and restoration techniques, especially related to soil water conditions.

Acknowledgments
We thank all sponsors (University of Zaragoza, IEA-DPH, DGA, MICCIN) and colleagues with whom we shared so greatly during this time the adventure of knowledge about the effect of fire on ecosystems.