**2. Erosion on contrasting soil types affected by fire**

Soil loss after fire in different soil types and plant community structures were analyzed in the semi-arid central Ebro Basin (NE-Spain) after fire.

## **2.1 Methods**

Soil loss after fire is affected by soil properties, plant community structure (i.e. seeders versus sprouters ratio) as well as fire and rainfall intensity. In this subchapter, we analyze

Soil Erosion and Conservations Measures in Semiarid Ecosystems Affected by Wildfires 89

Fig. 2. Erosion (g m-2) of the three types of burned soils (n=9). Rainfall simulator method.

relationship between erosion and plant cover was established after the fire.

during the first months (X, time) after a wildfire.

Valdejasa hills (NE-Spain).

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

 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)

Fig. 3. Accumulated soil erosion (g m-2) and bare soil (%) evolution in burned Castejón de

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.

#### **2.2 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).

Fig. 1. Fire effects on soil erosión (g m-2) in a Calcaric Regosol (rainfall simulations=9).

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 (Badía & Martí, 2003). 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.

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

Runoff and soil loss increased significantly after fire with exposure to strong rainfall (80 mm

Fig. 1. Fire effects on soil erosión (g m-2) in a Calcaric Regosol (rainfall simulations=9).

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 (Badía & Martí, 2003). 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

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.

(colluvial soils) as well as gypsiferous soils, classified as Haplic Gypsisol.

h-1) in a Calcaric Regosol below a *Pinus halepensis* forest (Fig.1).

developed in gypsum and colluvium (Fig. 2).

**2.2 Results and discussion** 

Fig. 2. Erosion (g m-2) of the three types of burned soils (n=9). Rainfall simulator method.

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.

$$\text{Sheet erosion (g m}^2\text{)} = 282 - 140 \,\text{.} \text{ (Plant cover, \text{\textquotesingle}\textquotesingle)} ; \text{r} = 0 ; 97 , \text{n} = 10 \tag{1}$$

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.

Fig. 3. Accumulated soil erosion (g m-2) and bare soil (%) evolution in burned Castejón de Valdejasa hills (NE-Spain).

Soil Erosion and Conservations Measures in Semiarid Ecosystems Affected by Wildfires 91

**(Kg ha-1 mm-1) Runoff (%) Other parameters** 

Tr=266-510 s

Sc=0,88-2,85 g L-1 ; Tr=186-360s

Sc=9,4-11,3 g L-1 ; Tr=260-360 s

EC=0,95-1,65 dS m-1

Tr=70-430 s

Sc=0,03-0,84g L-1

I=3,6-35,5 mm h-1 Wf=5-31 cm EC=1,0-2,5 dS m-1

Sc=0,46-1,06 g L-1 , Tr=67-665 s

Sc=0,43-0,97g L-1 , Tr=116-406 s

I=14-49 mm h-1 Sc=0,22-4,81 g L-1 EC=0,45-1,01 dS m-1 Tr=74-97 s ; Wf=10,4-25,2 cm

EC=0,62-0,91 dS m-1

EC=1,54-1,71 dS m-1

N-S aspect 0-68,2 0-73 I=15-55 mm h-1 ; Tr=138-872 s

N-S aspect 0-11,6 0-56 I=16-40 mm h-1

festilisers, crops 3,9-14,7 40-75 Wf=5-23 cm , Sc= 1,2-6,1 g L-1

treatments 5-105 5-95 Tr=53-253 s

soil moisture effect 0,2-8,6 2-55 I=17-49 mm h-1

soil moisture effect 0,2-8,6 0,1-29 I=33-42 mm h-1

(control vs burned) 8,5-18,8 24-25 I=44-45 mm h-1, Tr=197-379 s

(control vs burned) 25-152 22-60 I=24-47 mm h-1, Tr=199-232 s

Abbreviatures: I, Infiltration (mm h-1); Tr, Time to runoff (s); Sc, Sediment concentration (g L-1); Wf,

and season 24-374 36-85 EC=0,03-1,10 dS m-1

0,03-9,4 5-45 I=25 to 52 mm h-1

**Soil loss** 

1995 Plant cover 1,4-62,2 50-91 Sc=0-2-9,4 g L-1

1998 Plant cover - 24-50 I=24-37 mm h-1 , Wf=5-7 cm

1998 Plant cover 0,8-9,4 34-58 EC=0,13-0,16 dS m-1

1998 Plant cover 24,8-39,8 27-35 EC=0,28-0,29 dS m-1

<sup>2001</sup>Sever wildfire effect 3-76 23-45 - Cerdà, 2001 Rock fragments cover 0,36-14,1 12-38 I=27-44 mm h-1

<sup>2003</sup>Fire intensity effect 0,07-30,8 - -

raindrop size 0,73-39,9 41-81

Wetting front (cm); EC, Electrical conductivity of runoff (dS m-1) of overland flow.

Mediterranean ecosystems (arranged by publication date).

Table 1bis. Runoff and erosion data with rainfall simulator method in semi-arid

<sup>2004</sup>Slope gradient 1,4-23,0 34-58 Wf=5,2-8,5 cm

**Author Treatment or** 

Cerdà et al., 1995

Kutiel et al., 1995

Cerdà et al.,

Imeson et al.,

Imeson et al.,

Lasanta et al.,

Johansen et al.,

Cerdà & Doerr,

Calvo et al., 2003

Calvo et al., 2003

De Luis et al.,

Arnáez et al.,

Badía & Martí,

León et al., 2011 Fire effects

León et al., 2011 Fire effects

2008

2000

Ortiz & Alcañiz, 2001

2005

Cerdà & Lavee,

**variable** 

Fire, time and

Fire, time and

Abandonment,

Seeding and

Vegetation type, season and time from burning

N-S Aspect and

N-S Aspect and

Fire effects and

Desir, 2002 N-S Aspect 2,6-21,7 14-53

Cerdà, 2002 Parent material


Table 1. Runoff and erosion data with rainfall simulator method in semi-arid Mediterranean ecosystems (arranged by publication date).

sandy loam

Lithic Leptosol Limest. colluvium

Lithic Leptosol Limest. colluvium

300 mm 4-8 Loam Ponderosa

358 mm 36-58 Badlands Negligible

474 mm 40-50 Calcaric Regosol Pine forest &

387 mm 40-50 Lithic Leptosol Pine forest &

loamy

eroded

Regosols

Table 1. Runoff and erosion data with rainfall simulator method in semi-arid Mediterranean

560 mm 30-40 Haplic Gypsisol Shrubland 0,21 60'

sandy surface

& Luvisols

350 mm 25-45 Gypsisol, silty Shrubland 0,24 60'

690 mm 15-22 Chromic Luvisol Aleppo pine

**(%) Soil Vegetation** 

260 mm 35-40 Limestones Shrubland 0,24 60'

**Plot size (m2)**

shrubs 0,24 60'

forest 0,24 60'

Shrubland 0,24 60'

pine forest 32,4 120'

burned 0,24 40'

vegetation 0,24 60'

forest 0,24 60'

shrubland 0,24 60'

shrubland 0,24 60'

shrubland 4 105'

to shrubs 0,24 30'

forest 0,24 60'

and oak forest <sup>1</sup>120'

Pine &

Cambisol, silty Dehesa 0,24 60'

Aleppo pine

Gypsisols Barley, weeds 13,85 30'

Luvisols Shrubland 0,24 60'

Pines & oaks

Aleppo pine

Gorse

Bare soil

Aleppo pine

Phaeozem Shrubland 0,21 60'

**Rainfall duration intensity** 

55 mm h-1

30 mm h-1

37 mm h-1

50 mm h-1

50 mm h-1

50 mm h-1

60 mm h-1

60 mm h-1

55 mm h-1

95 mm h-1

55 mm h-1

55 mm h-1

45 mm h-1

55 mm h-1

55 mm h-1

156 mm h-1

75 mm h-1

85 mm h-1

60 mm h-1

60 mm h-1

**Author** 

Cerdà et al., 1995

Kutiel et al., 1995

Cerdà et al., 1998

Imeson et al.,

Imeson et al.,

Lasanta et al.,

Johansen et al., 2001

Ortiz & Alcañiz, 2001

Cerdà & Doerr, 2005

Calvo et al., 2003

Calvo et al., 2003

De Luis et al.,

Arnáez et al.,

2003

2004

Badía & Martí, 2008

León et al, 2011

León et al, 2011

Cerdà, 2001 Valencia

Cerdà, 2002 Alacant

Desir, 2002 Zaragoza

1998

1998

2000

Cerdà & Lavee, 1995

**Location and Mean Annual Precipitation** 

Valencia

Carmel Mont

Judea's Desert

Cáceres

Benidorm

Finestrat

Zaragoza

LosAlamos, USA

Taradell

Valencia

Callosa

Benidorm

Alicante

La Rioja

Fraga

Zuera

Zuera

ecosystems (arranged by publication date).

387 mm 36-40

400 mm 36-40

**Slope** 

688 mm 24-31 Calcareous,

511 mm 35-40 Dystric

324 mm <10 Calcisols,

688 mm 21-31 Leptosols &

563 mm 20-25 Luvisols,

688 mm 16-25 Lithic Leptosols

466 mm 50 Kastanozem,

800 mm 10-60 Kastanozem

318 mm 28-32 Calcaric

560 mm 30-40 Rendzic


Abbreviatures: I, Infiltration (mm h-1); 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).

Soil Erosion and Conservations Measures in Semiarid Ecosystems Affected by Wildfires 93

Fig. 4. Fire and drop size effects in the soil erosion (g m-2 h-1) with rainfall simulator method.

Fig. 5. Evolution of electrical conductivity of runoff (EC, dS m-1) in burned and control plots

with rainfall simulator method (in a Calcaric Regosol).

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 (Rodríguez et al., 2000). This is due to the differential evolution of vegetation to cover the soil surface after the fire.
