- Email: [email protected]
The differential response of surface runoff and sediment loss to wildfire events
Catena 121 (2014) 241–247
Contents lists available at ScienceDirect
Catena journal homepage: www.elsevier.com/locate/catena
The differential response of surface runoff and sediment loss to wildfire events Lea Wittenberg a,⁎, Dan Malkinson a,b, Ronel Barzilai a a b
Department of Geography and Environmental Studies, University of Haifa, Haifa 31905, Israel The Golan Research Institute, University of Haifa, Haifa 31905, Israel
a r t i c l e
i n f o
Article history: Received 25 July 2013 Received in revised form 29 April 2014 Accepted 20 May 2014 Available online xxxx Keywords: Forest fires Mediterranean ecosystems Soil erosion Surface runoff Vegetation recovery rates
a b s t r a c t Wildfires are of primary importance in determining ecosystem function and geomorphological process in most of the forested landscapes across the globe. Following a fire event in a maquis forest located at the Carmel mountain range in northern Israel, we monitored the eco-geomorphic response of the system, in an attempt to explain runoff and sediment yield dynamics. Specifically, we assessed growth of vegetation cover, and monitored runoff and sediment yield in relation to three controlling factors: fire severity, slope aspect, and slope steepness. Fourteen 10 m2 plots were constructed in different combinations of aspect, fire severity and steepness, which were monitored for a period of 24 months. Analysis of vegetation cover indicated that initial growth was faster on the north aspects, but by the end of the study period vegetation cover was similar to that of pre-fire levels on both aspects. Runoff and soil loss amounts from the burnt sites were commonly significantly higher on the south slope, steep gradients and high fire severity, compared to the counterpart plots. Temporal analysis indicated that sediment yield from the plots significantly decreased between the first and second winter seasons, whereas no statistically significant decrease in runoff was observed. Applying regression analysis methods we investigated the response of sediment yield to runoff, vegetation cover, soil moisture, rain intensity and precipitation, with respect to each of the controlling factors. In all cases runoff appeared to be a significant variable, as was vegetation cover, with the exception of the moderate burnt plots. We suggest that vegetation plays a complex role in determining the response of the geomorphic system to wildfire perturbations. While the mere presence of vegetation is sufficient to reduce soil loss, it is not sufficient to significantly affect runoff, most likely due to the different architecture of the newly regenerated vegetation. Additionally, vegetation seems to be an important factor in the harsher environments where more intensive soil movements occur, as the conditional effect of vegetation is more pronounced, and its contribution to the reduction of soil movements is higher. © 2014 Published by Elsevier B.V.
1. Introduction Forest fires play an important and prolonged role in structuring the complex eco-geomorphic systems of the Mediterranean basin. There is evidence that fires were frequent during the late Quaternary (Carrión et al., 2003), and probably even earlier, as many species have acquired adaptive mechanisms to endure and regenerate after recurrent fire events (Ne'eman et al., 2004; Pausas and Verdú, 2005). It is widely agreed that the immediate effects of wildfires on the chemical and physical properties of the soil–vegetation system (Certini, 2005; DeBano, 2000; González-Pérez et al., 2004; Wittenberg, 2012), coupled with reductions in biomass, facilitate erosive overland flow from the burnt sites, up to 5 orders of magnitudes higher than from natural, nonburnt rates (Inbar et al., 1998; Mayor et al., 2007; Rulli et al., 2006). ⁎ Corresponding author. E-mail address: [email protected] (L. Wittenberg).
http://dx.doi.org/10.1016/j.catena.2014.05.014 0341-8162/© 2014 Published by Elsevier B.V.
The uppermost soil loss occurs mostly within the first years, followed by a sharp decrease in erosion rates and generally, within the first decade geomorphic processes rates return to their pre-disturbance conditions (Gimeno-Garcia et al., 2007; Shakesby, 2011; Wittenberg and Inbar, 2009). Among the commonly considered factors in the literature determining post-fire hydrological and sedimentological dynamics, is the formation and destruction of hydrophobic layers (Debano, 2000; Doerr et al., 2000), fire-induced ash formation (Woods and Balfour, 2010), changes to the physical, chemical and biological soil properties (Mataix-Solera et al., 2011; Wittenberg, 2012), and damage to soil organic matter and the aboveground biomass (Doerr et al., 2009; Neary et al., 1999). Postfire soil losses observed in the Mediterranean are relatively variable and depend on vegetation composition and soil type, post-fire weather conditions and fire severity (Shakesby, 2011). Another key factor which plays a fundamental role affecting the hydrological processes is vegetation cover. The dominant mechanism by
242
L. Wittenberg et al. / Catena 121 (2014) 241–247
which vegetation affects these processes include 1) rain drops interception through the plant canopy and the litter, which reduces splash erosion (Domingo et al., 1998; Neary et al., 2005; Zuazo et al., 2004); 2) funneling the intercepted rain drops on to the stem of the plant to its base where the soil is more permeable, resulting in higher local infiltration rates and capacities (Martinez-Meza and Whitford, 1996; Zuazo et al., 2004); 3) increasing hydraulic roughness which decreases runoff flow velocity (Valentin and d'Herbes, 1999); and 4) organic matter produced by the vegetation increases soil aggregate stability (Neary et al., 2005). These processes form a source-sink mosaic, in which bare soil functions as a source patch, providing sediments available for transport, and vegetation cover form sink patches, trapping sediments and enriching it with nutrients (Puigedefabregas, 2005). As a result, runoff rates from forested areas in the Mediterranean are generally negligible and the consequent soil loss is minimal. Thus, the removal of vegetation cover following a fire event is an important factor dictating geomorphological response. Vegetation regeneration, however, may commence immediately depending on the species composition and timing of the fire. Several studies indicate that vegetation cover values are lower during the first wet season after the fire, characterized by 20%–45% vegetation cover, compared to a coverage of 50%–70% during the second rain season (Daly et al., 2004; Inbar et al., 1998; Mayor et al., 2007; Porporato et al., 2003). Runoff and soil loss yields significantly decrease after a threshold of vegetation cover is achieved. In the Mediterranean, a threshold of 30–40% vegetation cover is sufficient to significantly decrease soil loss and runoff yields (De Luis et al., 2001; Gimeno-Garcia et al., 2007; Loch, 2000; Puigedefabregas, 2005). The role of vegetation is also manifested via the physiographic properties of landscapes. Varying slope properties produce different runoff and soil loss yields, due to differences in aspect, steepness, lithology and vegetation type (Cerda et al., 1995; Mayor et al., 2007). Studies conducted in the Mediterranean basin found runoff values and soil loss yield after wildfires to be higher on the southern aspects than on northern ones (Cerda et al., 1995; Keizer et al., 2005; Wittenberg and Inbar, 2009), presumably due to higher vegetation regeneration rates on northern aspects. Fire severity also determines the abovementioned factors which affect geomorphic response, as it has been shown that 1) hydraulic roughness is lower in severely burnt areas compared to moderately burnt ones (Valentin and d'Herbes, 1999), 2) in severely burnt areas, the lack of litter exposes the soil to the effect of raindrops and the formation of a soil-sealed layer (Zuazo et al., 2004), and 3) the potential formation of the water-repellent hydrophobic layer may facilitate faster runoff generation and increasing soil loss rates (Doerr et al., 2000). Following low-severity fires sediment loss is usually low due to the remaining foliage covering the soil (Pausas et al., 2008). In Mediterranean type ecosystems several studies indicated that as fire severity increases, the correlation between rainfall and erosion becomes stronger. Consequently, soil loss and runoff rates are higher in severity compared to moderately burnt sites (Campo et al., 2006; Gimeno-Garcia et al., 2007). In spite of the abundant literature pertaining to post-fire geomorphic processes, relatively little attention has been devoted to the intra- and inter-season revegetation properties and their influence on runoff and sediment yields. Herein, we analyze at a resolution of a single precipitation event the properties of these two phenomena following a fire event in a burnt Mediterranean maquis. Further, we analyze the dynamics of these variables with respect to the recovery of the vegetation cover, and their differential response in relation to fire intensity and different physiographic controlling factors. 2. Study site Mt. Carmel is a distinctive mountain ridge steeply rising along the NW coast of Israel, with its highest peak at 546 m a.s.l. (Fig. 1). The
Fig. 1. The research area location, Mt. Carmel, Israel.
permeable lithology is composed of upper Cretaceous carbonate rocks, mainly limestone, dolomite, chalk, marl and local exposures of volcanic tuff. The dominant soil type in the studied plots is brown rendzina (Lithic Haploxeroll), overlying soft limestone, chalk and marls (Soil Survey Staff, 2006). These soils have the following characteristics: total calcium carbonate content of 32%, pH of 7.4 and EC of 0.90 dS m−1. Due to fire effects the upper top soil layer is highly disrupted and mixed with vegetation ash; the organic horizon (A), up to 5–7 cm grades downward into weakly structured C horizons (7–40 cm). The soil is clay-loamy from the surface (with 33–48% of clay in the first 7 cm) and clayey below down to 40 cm of depth (with 53–61% clay and 24–32% silt). Organic matter varies widely with fire severity (4.6–17.4%), exhibiting an average of 11.2% in the burnt plots, compared to 13.6% in the adjacent non-burnt soils. The vegetation of Mt. Carmel is a typical Mediterranean maquis (Malkinson and Wittenberg, 2007), ranges from dense mixed pine (Pinus halepensis) and oak (Quercus calliprinos) forests to more open and patchy mosaic-like tree-shrub formations (Kutiel and Naveh, 1987). The climate is characteristic Mediterranean, where a long hot and dry season prevails during May–October, and the precipitation is concentrated during the colder winter months, accounting for over 95% of the annual total. Mean annual precipitation in the region is 710 mm (Halfon, 2003). During April 2005 a wildfire consumed 154 ha of forest on the southern fringe of the city of Haifa, located in the northern part of the Carmel ridge. The fire was generally classified as a medium severity fire with some localities being consumed by high-severity fires (Tessler et al., 2008). 3. Methodology One month after the fire event, runoff and sediment collecting plots were constructed. Plots were located after considering three physiographic controlling factors: slope gradient (steep N 15° vs. moderate b 15°), slope aspect (north vs. south) and fire severity
L. Wittenberg et al. / Catena 121 (2014) 241–247
(low and high), forming seven experimental combinations (Table 1). Altogether 14 plots were constructed, averaging 10.3 m 2 in area (sd = 0.9), and in each experimental combination two plots were set up. The plots were delimited using plastic lawn edging, which were inserted to a depth of 20 cm into the soil. The plastic sheets were cemented to the ground in order to prevent leakage of runoff or sediments from the plot, and to prevent run-on from being erroneously estimated. Runoff and sediment were collected into 60 L buckets located at the downslope side of the plots. To assess local precipitation variation induced by aspect, two rain gauges were installed in the study area, on the north and south aspect each. Rainfall intensities were measured at the Haifa University meteorological station, located less than 1 km from the study site. Rainfall intensity was characterized using the following indices: I30 — highest intensity for 30 minute period, I60 — highest intensity for 60 minute period and I10 for 10 minute periods. Fire severity was estimated based on field evidence (Cocke et al., 2005). Runoff and soil loss measurements were recorded and sampled following each effective (sediment producing) rainstorm. During the events all runoff and sediments were funneled to, and collected within, the buckets. Soil samples were oven dried for 24 h at 104 °C, weighed and net sediment yield (g m− 2) was calculated. Runoff coefficient (RC), the percentage of precipitation that appears as runoff over the plot surface area (%), was also calculated per event per plot. In order to assess the changes in vegetation cover, vertical photographs of the plots were acquired monthly using a digital camera mounted on a six meter pole (Canon Rebel 8 megapixels). Images of the non-burnt (control) plots were acquired twice a year (winter and summer). The images were rectified to a base plot, using ArcMap software Version 9.1 (ESRI, 2009). As the plots were not located on horizontal planes, third order transformations were applied to correct topographical variation within the plot (max Root Mean Square Error = 0.01 cm). Each image was classified using the Feature Analyst module to three land cover classes: soil, vegetation and stones. Classification errors were assessed using 300 randomly selected pixels per image, which were visually compared to the original image. Due to the deviations of the response variables from normal distributions, and the inability to normalize their distributions, the nonparametric Mann–Whitney test statistics were applied. Additionally, we examined the contribution and significance of the different drivers which affect soil yield in relation to the controlling factors by applying stepwise multiple regression analyses (SAS, 2009). Specifically, we evaluated the role of precipitation, storm intensity, storm interval, vegetation cover, and runoff in relation to sediment yield. 4. Results 4.1. Vegetation regeneration and runoff The steepness of the moderate plots ranged between 8.9° and 14.5° (x = 11.7°, sd = 2.5) and that of steep plots between 17.2° and 24.8° (x = 20.8°, sd = 5.8). During the 2005/06 winter season, 526 mm were recorded at the Haifa university meteorological station, which amounted to 74% of the long-term mean annual precipitation. Eleven rain events were measured in the plots, during which 471 mm and 431 mm were recorded on the north and south aspects respectively, with maximum intensities of I30 — 23.8 mm/h, I60 — 12.5 mm/h and I10 — 8.3 mm/10 min. During the 2006/07 rain season, 667 mm were measured, which was 94% of the long-term average. In the course of the second winter season 13 rain events occurred and precipitation amounts on the north and south aspects were 601 and 595.8 mm, respectively. Maximum intensities of I30 — 20.4 mm/h, I60 — 12.5 mm/h and I10 — 5.7 mm/10 min were measured. A monthly time series of the vegetation cover was obtained by classifying the “aerial photographs” taken at each plot. Following image classification vegetation cover was estimated in terms of the percentage
243
Table 1 Research plots attributes: plot area, aspect, steepness and fire severity. Plot
Aspect
Fire severity
Gradient
Plot area (m2)
1a 1b 2a 2b 3a 3b 4a 4b 5a 5b 6a 6b 7a 7b
North North North North North North South south South South North North South South
Low Low High High Low Low High High Low Low High High High High
Moderate Moderate Steep Steep Steep Steep Steep Steep Steep Steep Moderate Moderate Moderate Moderate
10.41 9.50 12.48 9.30 9.72 10.50 11.73 10.40 10.35 10.34 10.90 10.04 10.39 9.10
of land cover occupied by it. During November 2005, five months following the fire, mean vegetation cover in the plots was 8.4% (sd = 6.0). During the first winter season, between November 2005 and February 2006, vegetation cover remained constant and no additional growth was observed (Fig. 2). During the last months of the first winter (2005), vegetation regeneration resumed and by March vegetation cover was 15.1% (sd = 7.2); two months later mean vegetation cover reached 28.6% (sd = 11.0). During the dry months (June–August 2005) vegetation cover remained constant on the south aspect plots, while on the north aspect vegetation cover decreased due to the withering of the annual vegetation. From September 2006 (mean vegetation cover 27%, sd = 12.25) to March 2007 (62%, sd = 12.07) regeneration was rapid, followed by a slight decrease in revegetation rates. By the end of the research period mean vegetation cover was 67.2% on the northern aspects, and 54% in the southern plots. These values are comparable with the control plots cover, where in January vegetation cover was 70.5% and 61.3% in the north and south aspects, respectively. Vegetation cover pertains merely to the aerial extent covered by the vegetation in the plot, and it does not imply that the vegetation's structural properties are similar to the control plots. Among all the controlling factors investigated (slope gradient, aspect and fire severity), differences in vegetation cover changes were associated only with aspect, where vegetation cover was significantly higher on the north aspects (t = 2.204, p = 0.028). Twenty four rain events were recorded during the study period, yielding a mean runoff ranging between 0.02 and 16.48 L m− 2 per event. The maximum runoff coefficients (RC) – 52% – occurred during the second rain event, in which both rain intensity (max I60 = 3.6 mm h−1) and precipitation amount (8.9 mm) were of the lowest values measured during the first season. Comparing runoff coefficients (RC) with respect to the controlling factors by year (Table 2) indicated that RCs were significantly higher on the south aspect and, steep gradient plots, but were not significantly affected by fire severity. In some cases differences in RC were higher by an order of magnitude as is the case in the north vs. south aspects during the second winter season. Pooled mean runoff per event and standard deviation during the first and second rain seasons are presented in Table 2. The results indicate that a general but insignificant decreasing trend of mean seasonal RC existed. The standard deviation also decreased when comparing the different controlling factors between the first and second winter seasons, suggesting some kind of a convergent response of the plots. 4.2. Sediment yield The overall mean sediment yield during the first winter was 42.2 g m− 2 per rain event, which markedly decreased during the second winter to 4.0 g m−2. The maximum sediment yield, 3436.8 g m−2, was generated in the first winter during a 126 mm rainstorm. To enable comparison among seasons and controlling factors, plots and rain events, we normalized the data per m2 and per mm of precipitation, generating a
244
L. Wittenberg et al. / Catena 121 (2014) 241–247
Fig. 2. Vegetation cover regeneration — north and south aspects, and mean vegetation cover during the 24 month study period.
sediment yield coefficient, which is essentially similar to the RC index. Similarly to the runoff, sediment yields during both rain seasons were significantly higher on the southern aspects compared to the northern ones (Table 3). Likewise, soil losses were significantly higher on steep slopes compared to moderate ones. To compare within-year differences, in response to the different controlling factors, the dataset was partitioned by winter season and each of the sub-sets was analyzed for each of the controlling factors. Considering the problems associated with multiple comparisons we used a conservative rejection level of α = 0.05 / 3 = 0.0166 to detect significant differences. Mean normalized soil losses exhibited a decreasing trend of an order of magnitude between the first (x = 0.71 g m−2 mm−1) and the second (x = 0.08 g m−2 mm−1) winter seasons. The strong decrease in sediment yield across the years and with respect to the different controlling factors was significantly different in all the categories compared. The highest decrease in soil loss amounts (per unit area, per mm of rainfall) between the first and second seasons occurred at the southern aspects and on the steep-slope plots, where absolute values decreased from 20.9 to 6.2 g m−2 mm−1 and from 16.9 to 5.4 g m−2 mm−1, respectively. In relative terms (i.e. proportional decrease), however, the most pronounced reduction occurred in the moderate fire-severity plots, where inter-season values differed by a factor of 25, representing a decrease from 1.78 to 0.19 g m−2 mm−1 (Table 3). The associated standard deviations also decreased between the first and second wet seasons, indicating lower fluctuations in soil loss amounts between events and plots during the second season (see Table 3).
4.3. Regression analyses — runoff and sediment yield Multiple regression analyses were applied to each of the controlling factors in order to analyze the soil loss mechanisms and identify the Table 2 Mean RC and standard deviation (sd) of differing physiographic attributes during the first and second winters. Δt significance denotes differences between the two seasons. Δ factor significance indicates differences between two counterpart sets of plots. 2005–2006
North aspect South aspect Δ factor significance Moderate slope Steep slope Δ factor significance Mod fire severity High fire severity Δ factor significance
2006–2007 sd
x RC (%)
sd
7.8 23.7 0.0008 11.8 16.6 0.0017 10.8 17.9 0.051
9.7 25.6
5.6 17.1 b0.0001 6.01 13.6 b0.0001 8.7 11.8 0.12
8.8 18.0
0.15 0.40
16.3 10.2
0.40 0.47
10.7 17.4
0.847 0.13
15.4 22.5
Table 3 Mean soil losses (g m−2 mm−1) and standard deviation (sd) on differing physiographic attributes during the first and second wet winters. Δt significance denotes differences between the two seasons. Δ factor significance indicates differences between two counterpart sets of plots. 2005–2006
Δt significance (p-val)
x RC (%)
19.5 19.7
significant parameters associated with this process (Table 4). Consequently, the role of five independent variables was assessed during the regression analyses: precipitation, runoff, precipitation days, 30-min storm intensity and vegetation cover. Precipitation reflects the event cumulative rainfall amounts. Runoff is the amount of running water collected following each rain event. ‘Precipitation days’ is defined as the amount of precipitation in a rain event previous to the measured one, divided by the number of dry days between the successive events, i.e., the number of dry days between the previous and measured runoff event (precip-days = mme − 1 / dry days(e − 1,e)). This index served as a proxy for soil moisture, which was included as an additional independent variable in the analysis. The 30-min storm intensity was based on the maximum measured rainfall amounts per 30 min during each event. This variable, however, was never found to be significant. The adjusted R2 of the regression models explained between 7 and 32.1% of the variation, and was highest in the second-winter model. The models suggest that the major mechanism controlling soil loss is the runoff, which appeared to be a significant factor in all the regression models. Vegetation cover and precipitation were significant factors in six of the nine analyses performed, while “precip-days” was a significant variable only in three models. The analysis of the entire dataset yielded three significant variables driving sediment yield: runoff, vegetation cover and precipitation. Comparing between the first and second winter seasons (Table 4) suggests that while runoff was a significant factor during both winters, its relative contribution to sediment yield was seven times higher during the first winter (β = 3.58, β = 0.5 second winter, β's being the parameter coefficients in the regression models). Further, during the first winter the precipitation amounts per se was a significant factor driving sediment
−2
x (g m North aspect South aspect Δ factor sig. Man–W-U Moderate slope Steep slope Δ factor sig. High fire severity Moderate fire Δ factor sig.
Δt sig.
2006–2007 −1
mm
)
sd
−2
x (g m
−1
mm
)
sd
0.6 20.9 0.006
1.55 5.74
0.06 2.2 b0.001
0.13 6.21
b0.001 0.03
0.45 16.9 b0.001 16.4 1.78 b0.001
1.33 51.8
0.05 1.7 b0.001 1.8 0.07 0.005
0.11 5.4
0.01 b0.001
5.7 0.19
0.001 0.01
51.9 11.2
L. Wittenberg et al. / Catena 121 (2014) 241–247
245
Table 4 Soil loss regression model components — β coefficient values.
1 2 3 4 5 6 7 8 9
Controlling factor
Runoff (L mm−1)
Vegetation cover (%)
Event precipitation
Precip.-days
Intercept
Adj. R2
df
Overall model North aspect South aspect Steep slope Moderate slope High fire severity Moderate fire First winter Second winter
1.55 2.28 2.27 21.5 1.28 21.86 10.1 3.58 0.50
−0.013 −0.0009 −0.033 −0.36 −0.009 −0.23 – – –
0.001 – 0.024 0.199 0.005 0.18 – 0.017 –
– – – – 0.09 – 0.51 – −0.005
0.05 0.04 −0.13 12.4 −0.37 −0.8 −4.46 −0.77 0.22
0.15 0.10 0.20 0.22 0.16 0.2 0.07 0.287 0.32
255 136 107 139 101 128 116 99 144
yield, while during the second one it was more closely associated with precipitation amounts. A comparison between the north and south aspects indicates that at the southern aspects the magnitude of the effect of runoff is almost similar to the northern ones. In contrast, the effect of runoff on sediment yield is almost 17 times higher on steep slopes (β = 21.5) compared to moderate slopes (β = 1.28), and twice as high at the severely burnt (β = 21.86) plots compared to the moderately burnt ones (β = 10.1). The models also show vegetation cover to be a dominant parameter negatively correlated with soil loss. A comparison between the coupled controlling factors indicates that the importance of vegetation (or lack of it) is 36 times higher at southern aspects (β = − 0.033) compared to the northern ones (β = −0.0009), and 40 times higher at the steep slopes (β = −0.36) compared to the moderate ones (β = −0.009). Precipitation also appeared to be an important parameter in determining soil loss, and its relative importance was always more prominent in what we consider the extreme conditions, i.e. at the steeper slopes, southern aspects, high severity burnt plots and during the first winter season (Table 4). We also assessed the relationship between runoff and sediment yield, normalized over the amount of precipitation per rain event, during the two rain seasons. During the first season the relationship between these two parameters was weak, as expressed by a lower R2 value of 0.158. During the second season this relationship was stronger, yielding an R2 value of 0.314 (Fig. 3). Each unit of runoff generated less sediment yield during the second season as can be observed by the moderate slope of the regression line during that season (1st year slope = 2.01, 2nd year slope = 0.499). This may suggest that as the relationship becomes stronger between runoff and sediment yield, similar explanatory variables drive both of them.
The vegetation recovery rates, measured in this study, were considerably lower during the first year (26.6%) compared to the second one (62.5% vegetation cover), during which it increased by ~ 36%. By the end of the study period, two years following the fire, vegetation cover had almost reached its pre-fire values. As stated above, it has been suggested that a threshold of 20%–45% coverage is required to considerably reduce runoff yields and soil loss from burnt plots (Inbar et al., 1998; Mayor et al., 2007; Poser and Williams, 1998). Dieckmann et al. (1992) indicated that a 30% vegetation cover reduces the sediment yield by 90% on slopes of less than 30%. Clearly, this threshold has been achieved and even exceeded, but resulted only in a slight, non-significant decrease in runoff yields between the first and second winter seasons. Runoff yields remained relatively high throughout the 2-year study period, even in cases where vegetation cover values were similar to those of the control plot. These findings contradict the commonly reported results which indicate a significant decrease in runoff and sediment yields following the first post-fire winter (Inbar et al., 1998; Puigedefabregas, 2005).
5. Discussion In this study, we established a factorial experimental design to investigate the interconnected effects of slope gradient, aspect and fire severity, and their interrelationships with vegetation recovery patterns, on runoff and erosion processes in a burnt Mediterranean maquis. Runoff and soil loss amounts were commonly significantly higher on the south aspect, on steep gradients and with high fire severity, compared to the north aspect, moderate slopes and moderate fire severity. These results are in accordance with previous fire-related studies conducted in Mediterranean type ecosystems (Campo et al., 2006; Cerda et al., 1995; Gimeno-Garcia et al., 2007; Keizer et al., 2005; Marques and Mora, 1992; Pierson et al., 2002). However, given the experimental design and the marked differences between the burnt plots, particular attention was given to the inconclusive role of plant cover in determining runoff and erosion processes at the particular conditions. Previous studies suggest that post-fire elevated rates of runoff and erosion decline exponentially as vegetation cover increases. A comprehensive review provided by Zuazo and Pleguezuelo (2009) demonstrated that the exponent coefficient values of post-fire elevated rates of runoff and erosion ranges from 0.01 to 0.08, suggesting similar decrease rates in both of these processes. The results of this study, however, do not fall in line with these observations.
Fig. 3. The relationship between runoff and sediment yield during the two winter seasons: a) 2005–6 winter, b) 2006–7 winter. Note the differences in y-axis scale.
246
L. Wittenberg et al. / Catena 121 (2014) 241–247
Presumably, following forest fires, vegetation cover alone is not sufficient to control post-fire hydrological and sedimentological processes and additional properties of the vegetation, such as type and architecture (Zuazo et al., 2004), as well as physio-chemical changes of the soil, play a critical role in determining runoff patterns. As noted in the Results section, RC did not significantly decrease between the two winter seasons, whereas within each year significant differences were observed between the north and south aspects, and the moderate and steep slopes. Variations in fire severity, however, were not sufficient to generate significant differences in runoff. These results suggest that in the case of the Carmel 2005 fire vegetation and topography were key factors in determining post-fire runoff patterns, compared to variation in fire severity which did not generate such a response. In the light of the above discrepancies in runoff and erosion rates and to better understand the specific role of the intra- and inter-season soil yield generating mechanisms, multiple regression analyses were conducted (Table 4). The most influential parameter, which was positively correlated with soil loss, was the runoff. However, runoff did not significantly decrease between the first and second winter seasons, thus it cannot be held accountable for the inter-season decrease in soil loss. We assume that runoff, which depends on the coupled effects of event-specific precipitation properties and soil–vegetation condition, determines soil loss trends which occurred during each of the two rainy seasons. Vegetation cover also emerged as a significant factor driving sediment yield. Regression results point to the differential importance of the vegetation in the “extreme” conditions (steep slopes, severe fires and southern aspect) compared to the moderate ones. In the extreme conditions the absolute values of the vegetation β coefficients were much higher. This suggests that any additional unit of vegetation cover results in larger reductions of sediment yield. Thus we suggest that in “extreme” landscapes geomorphic sensitivity to fire disturbance is moderated to a larger extent by vegetation recovery rates and properties. While soil loss exhibited a decreasing trend during the study period, runoff did not display a similar response. Possibly, these patterns might be explained by the differential effect that vegetation has on these processes. For example, during the first winter season a RC of 2.1% produced 20.5 g m−2, while during the second winter a higher RC (3.6%) produced only 5.1 g m−2. This might indicate that vegetation cover values are sufficiently high to moderate soil loss processes. These responses should also be interpreted in light of the results presented in Fig. 3, which indicate that during the second winter runoff produces less soil loss, as expressed by the slope of the lines, but a higher proportion of the variation in soil loss is accounted for by runoff. Thus a complex set of interactions emerged in this system: runoff did not respond to changes in vegetation cover, whereas soil loss did, and at the same time the correlation between runoff and soil loss increased. This may suggest that other factors which affected runoff and soil loss during the first winter season became less dominant during the second one. Such factors may include presence of ash and hydrophobicity which have been observed to have relatively short term effect Bodí et al. (2012). Alternatively, soil resources could have been exhausted during the first winter, resulting in lower loss values during the following winter. Soil loss may be directly affected by canopy regeneration, whereas runoff is hypothesized to be determined by plant architecture, such as plant canopy and stem structure (Domingo et al., 1998; Martinez-Meza and Whitford, 1996; Zuazo et al., 2004), and possibly the spatial pattern of the vegetation (Mayor et al., 2008). Therefore, a sorting pattern emerges, in which soil loss dynamics change at a certain rate in response to the disturbance event, while runoff yields respond at a different rate. This allows addressing two central aspects; the first, a scientific geomorphological approach, is aimed to better understand runoff and soil loss processes, in the light of vegetation recovery dynamics, while the other is an applied approach, aiming to locate vulnerable areas during the rehabilitation process, in order to minimize the damage inflicted by wildfires.
6. Conclusions The resilience of an eco-geomorphic system, namely, its ability to absorb a perturbation and to recover to its pre-disturbance structure and state, is a vital intrinsic property of the system (Westman, 1986). Mediterranean ecosystems are largely considered as being resilient to fire, soil disturbances and grazing (Lavorel, 1999). The resilience to fire, however, may fundamentally differ within the disturbed area and among the various patches of the burnt landscape, given the specific biotic and physiographic interactions occurring within each of the areas. Essentially, this study was concerned with the geomorphic response of perturbed geo-systems characterized by different physiographic conditions. Runoff and sediment yield from the burnt plots responded differentially to the increase in post-fire vegetation cover, where runoff response seems to lag behind soil loss. Two years after the fire vegetation cover was almost completely restored in terms of areal cover, although not necessarily with respect to its structure. Nonetheless, we suggest that the mere increase in vegetation cover is not sufficient to significantly affect reduction in runoff, and additional properties of the vegetation ought to be considered, assumingly plant architecture and its spatial distribution. Further, the results indicate that the increase in vegetation cover has a more pronounced effect in locations characterized by higher geomorphic sensitivity (sensu Swanson, 1981).
References Bodí, M.B., Doerr, S.H., Cerdà, A., Mataix-Solera, J., 2012. Hydrological effects of a layer of vegetation ash on underlying wettable and water repellent soil. Geoderma 191, 14–23. Campo, J., Andreu, V., Gimeno-Garcia, E., Gonzalez, O., Rubio, J., 2006. Occurrence of soil erosion after repeated experimental fires in a Mediterranean environment. Geomorphology 82, 376–387. Carrión, J.S., Sanchez-Gomez, P., Mota, J.F., Yll, R., Chain, C., 2003. Holocene vegetation dynamics, fire and grazing in the Sierra de Gador, southern Spain. The Holocene 13, 839–849. Cerda, A., Imeson, A., Calvo, A., 1995. Fire and aspect induced differences on the erodibility and hydrology of soils a La Costera, Valencia, Southern Spain. Catena 24, 289–304. Certini, G., 2005. Effects of fire on properties of forest soils: a review. Oecologia 143, 1–10. Cocke, A.E., Fulé, P.Z., Crouse, J.E., 2005. Comparison of burn severity assessments using Differenced Normalized Burn Ratio and ground data. Int. J. Wildland Fire 14, 189–198. Daly, E., Porporato, A., Rodrigez-Iturbe, I., 2004. Coupled dynamics of photosynthesis, transpiration and soil water balance I. Upscaling from hourly to daily time scale. J. Hydrometeorol. 5, 546–558. De Luis, M., Garcia-Cano, M.F., Cortina, J., Raventos, J., Gonzalez-Hidalgo, C.J., Sa'nchez, J.R., 2001. Climatic trends, disturbances and short-term vegetation dynamics in a Mediterranean shrubland. For. Ecol. Manag. 147, 25–37. DeBano, L.F., 2000. Water repellency in soils: a historical overview. J. Hydrol. 231–232, 4–32. Dieckmann, H., Motzer, H., Harres, H.P., Seuffert, O., 1992. Vegetation and erosion. Investigation on erosion plots in Southern Sardinia. Geooko-plus 3, 139–149. Doerr, S.H., Shakesby, R.S., Walsh, R.P.D., 2000. Soil water repellency: its characteristics, causes and hydro-geomorphological consequences. Earth-Sci. Rev. 51, 33–65. Doerr, S.H., Woods, S.W., Martin, D.A., Casimiro, M., 2009. Natural background soil water repellency in conifer forests of the north-western USA: its prediction and relationship to wildfire occurrence. J. Hydrol. 371, 12–21. Domingo, F., Sanchez, G., Moro, M., Brenner, A., Puigdefabregas, J., 1998. Measurement and modeling of rainfall interception by three semi-arid canopies. Agric. Meteorol. 91, 272–292. ESRI, 2009. ArcGIS. Environmental Systems Resource Institute, Redlands, CA, USA. Gimeno-Garcia, E., Andreu, V., Rubio, J., 2007. Influence of vegetation recovery on water erosion at short and medium-term after experimental fires in a Mediterranean shrubland. Catena 69, 150–160. González-Pérez, J.A., González-Vila, F.J., Almendros, G., Knicker, H., 2004. The effect of fire on soil organic matter — a review. Environ. Int. 30 (6), 855–870. Halfon, N., 2003. Translocation of the orographic rainfall enhancement from the ridge towards the lee slopes in northern Israel, unpublished M.Sc. thesis. Haifa University, Israel. p. 145 (in Hebrew). Inbar, M., Tamir, M., Wittenberg, L., 1998. Runoff and erosion processes after a forest fire in Mount Carmel, a Mediterranean area. Geomorphology 24, 17–33. Keizer, J.J., Coelho, C.O.A., Shakesby, R.A., Domingues, C.S.P., Malvar, M.C., Perez, I.M.B., Matias, M.J.S., Ferreira, A.J.D., 2005. The role of soil water repellency in overland flow generation in pine and eucalypt forest stands in coastal Portugal. Aust. J. Soil Res. 43, 337–349. Kutiel, P., Naveh, Z., 1987. The effect of fire on nutrients in a pine forest soil. Plant Soil 104, 269–274. Lavorel, S., 1999. Ecological diversity and resilience of Mediterranean vegetation to disturbance. Diversity and Distributions 5 (1–2), 3–13.
L. Wittenberg et al. / Catena 121 (2014) 241–247 Loch, R., 2000. Effects of vegetation cover on runoff and erosion under simulated rain and overland flow in a rehabilitated site on the Meandu Mine, Tarong, Queensland. Aust. J. Soil Res. 38, 299–312. Malkinson, D., Wittenberg, L., 2007. Scaling the effects of riparian vegetation on crosssectional characteristics of ephemeral mountain streams — a case study of Nahal Oren, Mt. Carmel, Israel. Catena 69, 103–110. Marques, M., Mora, E., 1992. The influence of aspect on runoff and soil loss in a Mediterranean burnt forest (Spain). Catena 19, 333–344. Martinez-Meza, E., Whitford, W., 1996. Stemflow, throughfall and channelization of stemflow by roots in three Chihuahuan desert shrubs. J. Arid Environ. 32, 271–287. Mataix-Solera, J., Cerdà, A., Arcenegui, V., Jordán, A., Zavala, L.M., 2011. Fire effects on soil aggregation: a review. Earth Sci. Rev. 109, 44–60. Mayor, A.G., Bautista, S., Llovet, J., Bellot, J., 2007. Post-fire hydrological and erosional responses of a Mediterranean landscape: seven years of catchment-scale dynamic. Catena 71, 68–75. Mayor, A.G., Bautista, S., Small, E.E., Dixon, M., Bellot, J., 2008. Measurement of the connectivity of runoff source areas as determined by vegetation pattern and topography: a tool for assessing potential water and soil losses in drylands. Water Resour. Res. 44, w10423. Neary, D.G., Klopatek, C.C., DeBano, L.F., 1999. Fire effects on belowground sustainability: a review and synthesis. For. Ecol. Manag. 122, 51–71. Neary, D.G., Ryan, K.C., DeBano, L.F., 2005. Wildland Fire in Ecosystems, Effects of Fire on Soil and Water. General Technical Report RMRS-GTR-42 — Volume 4 USDA, Rocky Mountain Research Station (252 pp.). Ne'eman, G., Goubitz, S., Nathan, R., 2004. Reproductive traits of Pinus halepensis in the light of fire — a critical review. Plant Ecol. 171, 69–79. Pausas, J., Verdú, M., 2005. Plant persistence traits in fire-prone ecosystems of the Mediterranean Basin: a phylogenetic approach. Oikos 109, 196–202. Pausas, J.G., Llovet, J., Rodrigo, A., Vallejo, R., 2008. Are wildfires a disaster in the Mediterranean basin? — A review. Int. J. Wildland Fire 17, 713–723. Pierson, F., Carlson, D., Spaeth, K., 2002. Impacts of wildfire on soil hydrological properties of steep sagebrush-steppe rangeland. Int. J. Wildland Fire 11, 145–151. Porporato, A., Do'dorico, P., Laio, F., Rodrigez-Iturbe, I., 2003. Hydrologic controls on soil carbon and nitrogen cycles. I. Modeling scheme. Adv. Res. Water Resour. 26, 45–58.
247
Poser, I.P., Williams, L., 1998. The effect of wildfire on runoff and erosion in native Eucalyptus forest. Hydrol. Process. 12, 251–265. Puigedefabregas, J., 2005. The role of vegetation patterns in structuring runoff and sediment fluxes in dryland. Earth Surf. Process. Landf 30, 133–147. Rulli, M.C., Bozzi, S., Spada, M., Bocchiola, D., Rosso, R., 2006. Rainfall simulations on a fire disturbed Mediterranean area. J. Hydrol. 327, 323–338. 2007SAS Institute Inc., 2009. JMP. v. 9. Cary, North Carolina, USA. Shakesby, R.A., 2011. Post-wildfire soil erosion in the Mediterranean: review and future research directions. Earth Sci. Rev. 105, 71–100. Soil Survey Staff, 2006. Keys to Soil Taxonomy, 10th ed. NRCS, Washington, DC. Swanson, F.J., 1981. Fire and geomorphic processes. In: Mooney, H.A., Bonnicksen, T.M., Christiansen, N.L., Lotan, J.E., Reiners, W.A. (Eds.), Fire Regime and Ecosystem Properties, United States Department of Agriculture, Forest Service, General Technical Report WO. United States Government Planning Office, Washington, DC, 26, pp. 401–421. Tessler, N., Wittenberg, L., Malkinson, D., Greenbaum, N., 2008. Short term changes in soil water repellency following a forest fire in Mount Carmel, Israel. Catena 74, 185–191. Valentin, C.D., d'Herbes, J.M., 1999. Niger bush as a natural water harvesting system. Catena 37, 1–24. Westman, W.E., 1986. Resilience: concepts and measures. in Resilience in mediterraneantype ecosystems (ed. by B. Dell, A.J.M. Hopkins & B.B. Lamont), pp: 5–19. Wittenberg, L., 2012. Post-fire soil ecology: properties and erosion dynamics. Isr. J. Ecol. Evol. 58 (2), 151–164. Wittenberg, L., Inbar, M., 2009. The role of fire disturbance on runoff and erosion processes — a long-term approach, Mt. Carmel case study, Israel. Geogr. Res. 47, 46–56. Woods, S.W., Balfour, V.N., 2010. The effects of soil texture and ash thickness on the postfire hydrological response from ash-covered soils. J. Hydrol. 3933, 274–286. Zuazo, V.D., Pleguezuelo, C.R.R., 2009. Soil-Erosion and Runoff Prevention by Plant Covers: A Review. Sustainable Agriculture. Springer, Netherlands, pp. 785–811. Zuazo, V.D., Martinez, J.F., Raya, A., 2004. Impact of vegetation cover on runoff and soil erosion at hillslope scale in Lanjaron, Spain. Environmentalist 24, 39–48.
Contents lists available at ScienceDirect
Catena journal homepage: www.elsevier.com/locate/catena
The differential response of surface runoff and sediment loss to wildfire events Lea Wittenberg a,⁎, Dan Malkinson a,b, Ronel Barzilai a a b
Department of Geography and Environmental Studies, University of Haifa, Haifa 31905, Israel The Golan Research Institute, University of Haifa, Haifa 31905, Israel
a r t i c l e
i n f o
Article history: Received 25 July 2013 Received in revised form 29 April 2014 Accepted 20 May 2014 Available online xxxx Keywords: Forest fires Mediterranean ecosystems Soil erosion Surface runoff Vegetation recovery rates
a b s t r a c t Wildfires are of primary importance in determining ecosystem function and geomorphological process in most of the forested landscapes across the globe. Following a fire event in a maquis forest located at the Carmel mountain range in northern Israel, we monitored the eco-geomorphic response of the system, in an attempt to explain runoff and sediment yield dynamics. Specifically, we assessed growth of vegetation cover, and monitored runoff and sediment yield in relation to three controlling factors: fire severity, slope aspect, and slope steepness. Fourteen 10 m2 plots were constructed in different combinations of aspect, fire severity and steepness, which were monitored for a period of 24 months. Analysis of vegetation cover indicated that initial growth was faster on the north aspects, but by the end of the study period vegetation cover was similar to that of pre-fire levels on both aspects. Runoff and soil loss amounts from the burnt sites were commonly significantly higher on the south slope, steep gradients and high fire severity, compared to the counterpart plots. Temporal analysis indicated that sediment yield from the plots significantly decreased between the first and second winter seasons, whereas no statistically significant decrease in runoff was observed. Applying regression analysis methods we investigated the response of sediment yield to runoff, vegetation cover, soil moisture, rain intensity and precipitation, with respect to each of the controlling factors. In all cases runoff appeared to be a significant variable, as was vegetation cover, with the exception of the moderate burnt plots. We suggest that vegetation plays a complex role in determining the response of the geomorphic system to wildfire perturbations. While the mere presence of vegetation is sufficient to reduce soil loss, it is not sufficient to significantly affect runoff, most likely due to the different architecture of the newly regenerated vegetation. Additionally, vegetation seems to be an important factor in the harsher environments where more intensive soil movements occur, as the conditional effect of vegetation is more pronounced, and its contribution to the reduction of soil movements is higher. © 2014 Published by Elsevier B.V.
1. Introduction Forest fires play an important and prolonged role in structuring the complex eco-geomorphic systems of the Mediterranean basin. There is evidence that fires were frequent during the late Quaternary (Carrión et al., 2003), and probably even earlier, as many species have acquired adaptive mechanisms to endure and regenerate after recurrent fire events (Ne'eman et al., 2004; Pausas and Verdú, 2005). It is widely agreed that the immediate effects of wildfires on the chemical and physical properties of the soil–vegetation system (Certini, 2005; DeBano, 2000; González-Pérez et al., 2004; Wittenberg, 2012), coupled with reductions in biomass, facilitate erosive overland flow from the burnt sites, up to 5 orders of magnitudes higher than from natural, nonburnt rates (Inbar et al., 1998; Mayor et al., 2007; Rulli et al., 2006). ⁎ Corresponding author. E-mail address: [email protected] (L. Wittenberg).
http://dx.doi.org/10.1016/j.catena.2014.05.014 0341-8162/© 2014 Published by Elsevier B.V.
The uppermost soil loss occurs mostly within the first years, followed by a sharp decrease in erosion rates and generally, within the first decade geomorphic processes rates return to their pre-disturbance conditions (Gimeno-Garcia et al., 2007; Shakesby, 2011; Wittenberg and Inbar, 2009). Among the commonly considered factors in the literature determining post-fire hydrological and sedimentological dynamics, is the formation and destruction of hydrophobic layers (Debano, 2000; Doerr et al., 2000), fire-induced ash formation (Woods and Balfour, 2010), changes to the physical, chemical and biological soil properties (Mataix-Solera et al., 2011; Wittenberg, 2012), and damage to soil organic matter and the aboveground biomass (Doerr et al., 2009; Neary et al., 1999). Postfire soil losses observed in the Mediterranean are relatively variable and depend on vegetation composition and soil type, post-fire weather conditions and fire severity (Shakesby, 2011). Another key factor which plays a fundamental role affecting the hydrological processes is vegetation cover. The dominant mechanism by
242
L. Wittenberg et al. / Catena 121 (2014) 241–247
which vegetation affects these processes include 1) rain drops interception through the plant canopy and the litter, which reduces splash erosion (Domingo et al., 1998; Neary et al., 2005; Zuazo et al., 2004); 2) funneling the intercepted rain drops on to the stem of the plant to its base where the soil is more permeable, resulting in higher local infiltration rates and capacities (Martinez-Meza and Whitford, 1996; Zuazo et al., 2004); 3) increasing hydraulic roughness which decreases runoff flow velocity (Valentin and d'Herbes, 1999); and 4) organic matter produced by the vegetation increases soil aggregate stability (Neary et al., 2005). These processes form a source-sink mosaic, in which bare soil functions as a source patch, providing sediments available for transport, and vegetation cover form sink patches, trapping sediments and enriching it with nutrients (Puigedefabregas, 2005). As a result, runoff rates from forested areas in the Mediterranean are generally negligible and the consequent soil loss is minimal. Thus, the removal of vegetation cover following a fire event is an important factor dictating geomorphological response. Vegetation regeneration, however, may commence immediately depending on the species composition and timing of the fire. Several studies indicate that vegetation cover values are lower during the first wet season after the fire, characterized by 20%–45% vegetation cover, compared to a coverage of 50%–70% during the second rain season (Daly et al., 2004; Inbar et al., 1998; Mayor et al., 2007; Porporato et al., 2003). Runoff and soil loss yields significantly decrease after a threshold of vegetation cover is achieved. In the Mediterranean, a threshold of 30–40% vegetation cover is sufficient to significantly decrease soil loss and runoff yields (De Luis et al., 2001; Gimeno-Garcia et al., 2007; Loch, 2000; Puigedefabregas, 2005). The role of vegetation is also manifested via the physiographic properties of landscapes. Varying slope properties produce different runoff and soil loss yields, due to differences in aspect, steepness, lithology and vegetation type (Cerda et al., 1995; Mayor et al., 2007). Studies conducted in the Mediterranean basin found runoff values and soil loss yield after wildfires to be higher on the southern aspects than on northern ones (Cerda et al., 1995; Keizer et al., 2005; Wittenberg and Inbar, 2009), presumably due to higher vegetation regeneration rates on northern aspects. Fire severity also determines the abovementioned factors which affect geomorphic response, as it has been shown that 1) hydraulic roughness is lower in severely burnt areas compared to moderately burnt ones (Valentin and d'Herbes, 1999), 2) in severely burnt areas, the lack of litter exposes the soil to the effect of raindrops and the formation of a soil-sealed layer (Zuazo et al., 2004), and 3) the potential formation of the water-repellent hydrophobic layer may facilitate faster runoff generation and increasing soil loss rates (Doerr et al., 2000). Following low-severity fires sediment loss is usually low due to the remaining foliage covering the soil (Pausas et al., 2008). In Mediterranean type ecosystems several studies indicated that as fire severity increases, the correlation between rainfall and erosion becomes stronger. Consequently, soil loss and runoff rates are higher in severity compared to moderately burnt sites (Campo et al., 2006; Gimeno-Garcia et al., 2007). In spite of the abundant literature pertaining to post-fire geomorphic processes, relatively little attention has been devoted to the intra- and inter-season revegetation properties and their influence on runoff and sediment yields. Herein, we analyze at a resolution of a single precipitation event the properties of these two phenomena following a fire event in a burnt Mediterranean maquis. Further, we analyze the dynamics of these variables with respect to the recovery of the vegetation cover, and their differential response in relation to fire intensity and different physiographic controlling factors. 2. Study site Mt. Carmel is a distinctive mountain ridge steeply rising along the NW coast of Israel, with its highest peak at 546 m a.s.l. (Fig. 1). The
Fig. 1. The research area location, Mt. Carmel, Israel.
permeable lithology is composed of upper Cretaceous carbonate rocks, mainly limestone, dolomite, chalk, marl and local exposures of volcanic tuff. The dominant soil type in the studied plots is brown rendzina (Lithic Haploxeroll), overlying soft limestone, chalk and marls (Soil Survey Staff, 2006). These soils have the following characteristics: total calcium carbonate content of 32%, pH of 7.4 and EC of 0.90 dS m−1. Due to fire effects the upper top soil layer is highly disrupted and mixed with vegetation ash; the organic horizon (A), up to 5–7 cm grades downward into weakly structured C horizons (7–40 cm). The soil is clay-loamy from the surface (with 33–48% of clay in the first 7 cm) and clayey below down to 40 cm of depth (with 53–61% clay and 24–32% silt). Organic matter varies widely with fire severity (4.6–17.4%), exhibiting an average of 11.2% in the burnt plots, compared to 13.6% in the adjacent non-burnt soils. The vegetation of Mt. Carmel is a typical Mediterranean maquis (Malkinson and Wittenberg, 2007), ranges from dense mixed pine (Pinus halepensis) and oak (Quercus calliprinos) forests to more open and patchy mosaic-like tree-shrub formations (Kutiel and Naveh, 1987). The climate is characteristic Mediterranean, where a long hot and dry season prevails during May–October, and the precipitation is concentrated during the colder winter months, accounting for over 95% of the annual total. Mean annual precipitation in the region is 710 mm (Halfon, 2003). During April 2005 a wildfire consumed 154 ha of forest on the southern fringe of the city of Haifa, located in the northern part of the Carmel ridge. The fire was generally classified as a medium severity fire with some localities being consumed by high-severity fires (Tessler et al., 2008). 3. Methodology One month after the fire event, runoff and sediment collecting plots were constructed. Plots were located after considering three physiographic controlling factors: slope gradient (steep N 15° vs. moderate b 15°), slope aspect (north vs. south) and fire severity
L. Wittenberg et al. / Catena 121 (2014) 241–247
(low and high), forming seven experimental combinations (Table 1). Altogether 14 plots were constructed, averaging 10.3 m 2 in area (sd = 0.9), and in each experimental combination two plots were set up. The plots were delimited using plastic lawn edging, which were inserted to a depth of 20 cm into the soil. The plastic sheets were cemented to the ground in order to prevent leakage of runoff or sediments from the plot, and to prevent run-on from being erroneously estimated. Runoff and sediment were collected into 60 L buckets located at the downslope side of the plots. To assess local precipitation variation induced by aspect, two rain gauges were installed in the study area, on the north and south aspect each. Rainfall intensities were measured at the Haifa University meteorological station, located less than 1 km from the study site. Rainfall intensity was characterized using the following indices: I30 — highest intensity for 30 minute period, I60 — highest intensity for 60 minute period and I10 for 10 minute periods. Fire severity was estimated based on field evidence (Cocke et al., 2005). Runoff and soil loss measurements were recorded and sampled following each effective (sediment producing) rainstorm. During the events all runoff and sediments were funneled to, and collected within, the buckets. Soil samples were oven dried for 24 h at 104 °C, weighed and net sediment yield (g m− 2) was calculated. Runoff coefficient (RC), the percentage of precipitation that appears as runoff over the plot surface area (%), was also calculated per event per plot. In order to assess the changes in vegetation cover, vertical photographs of the plots were acquired monthly using a digital camera mounted on a six meter pole (Canon Rebel 8 megapixels). Images of the non-burnt (control) plots were acquired twice a year (winter and summer). The images were rectified to a base plot, using ArcMap software Version 9.1 (ESRI, 2009). As the plots were not located on horizontal planes, third order transformations were applied to correct topographical variation within the plot (max Root Mean Square Error = 0.01 cm). Each image was classified using the Feature Analyst module to three land cover classes: soil, vegetation and stones. Classification errors were assessed using 300 randomly selected pixels per image, which were visually compared to the original image. Due to the deviations of the response variables from normal distributions, and the inability to normalize their distributions, the nonparametric Mann–Whitney test statistics were applied. Additionally, we examined the contribution and significance of the different drivers which affect soil yield in relation to the controlling factors by applying stepwise multiple regression analyses (SAS, 2009). Specifically, we evaluated the role of precipitation, storm intensity, storm interval, vegetation cover, and runoff in relation to sediment yield. 4. Results 4.1. Vegetation regeneration and runoff The steepness of the moderate plots ranged between 8.9° and 14.5° (x = 11.7°, sd = 2.5) and that of steep plots between 17.2° and 24.8° (x = 20.8°, sd = 5.8). During the 2005/06 winter season, 526 mm were recorded at the Haifa university meteorological station, which amounted to 74% of the long-term mean annual precipitation. Eleven rain events were measured in the plots, during which 471 mm and 431 mm were recorded on the north and south aspects respectively, with maximum intensities of I30 — 23.8 mm/h, I60 — 12.5 mm/h and I10 — 8.3 mm/10 min. During the 2006/07 rain season, 667 mm were measured, which was 94% of the long-term average. In the course of the second winter season 13 rain events occurred and precipitation amounts on the north and south aspects were 601 and 595.8 mm, respectively. Maximum intensities of I30 — 20.4 mm/h, I60 — 12.5 mm/h and I10 — 5.7 mm/10 min were measured. A monthly time series of the vegetation cover was obtained by classifying the “aerial photographs” taken at each plot. Following image classification vegetation cover was estimated in terms of the percentage
243
Table 1 Research plots attributes: plot area, aspect, steepness and fire severity. Plot
Aspect
Fire severity
Gradient
Plot area (m2)
1a 1b 2a 2b 3a 3b 4a 4b 5a 5b 6a 6b 7a 7b
North North North North North North South south South South North North South South
Low Low High High Low Low High High Low Low High High High High
Moderate Moderate Steep Steep Steep Steep Steep Steep Steep Steep Moderate Moderate Moderate Moderate
10.41 9.50 12.48 9.30 9.72 10.50 11.73 10.40 10.35 10.34 10.90 10.04 10.39 9.10
of land cover occupied by it. During November 2005, five months following the fire, mean vegetation cover in the plots was 8.4% (sd = 6.0). During the first winter season, between November 2005 and February 2006, vegetation cover remained constant and no additional growth was observed (Fig. 2). During the last months of the first winter (2005), vegetation regeneration resumed and by March vegetation cover was 15.1% (sd = 7.2); two months later mean vegetation cover reached 28.6% (sd = 11.0). During the dry months (June–August 2005) vegetation cover remained constant on the south aspect plots, while on the north aspect vegetation cover decreased due to the withering of the annual vegetation. From September 2006 (mean vegetation cover 27%, sd = 12.25) to March 2007 (62%, sd = 12.07) regeneration was rapid, followed by a slight decrease in revegetation rates. By the end of the research period mean vegetation cover was 67.2% on the northern aspects, and 54% in the southern plots. These values are comparable with the control plots cover, where in January vegetation cover was 70.5% and 61.3% in the north and south aspects, respectively. Vegetation cover pertains merely to the aerial extent covered by the vegetation in the plot, and it does not imply that the vegetation's structural properties are similar to the control plots. Among all the controlling factors investigated (slope gradient, aspect and fire severity), differences in vegetation cover changes were associated only with aspect, where vegetation cover was significantly higher on the north aspects (t = 2.204, p = 0.028). Twenty four rain events were recorded during the study period, yielding a mean runoff ranging between 0.02 and 16.48 L m− 2 per event. The maximum runoff coefficients (RC) – 52% – occurred during the second rain event, in which both rain intensity (max I60 = 3.6 mm h−1) and precipitation amount (8.9 mm) were of the lowest values measured during the first season. Comparing runoff coefficients (RC) with respect to the controlling factors by year (Table 2) indicated that RCs were significantly higher on the south aspect and, steep gradient plots, but were not significantly affected by fire severity. In some cases differences in RC were higher by an order of magnitude as is the case in the north vs. south aspects during the second winter season. Pooled mean runoff per event and standard deviation during the first and second rain seasons are presented in Table 2. The results indicate that a general but insignificant decreasing trend of mean seasonal RC existed. The standard deviation also decreased when comparing the different controlling factors between the first and second winter seasons, suggesting some kind of a convergent response of the plots. 4.2. Sediment yield The overall mean sediment yield during the first winter was 42.2 g m− 2 per rain event, which markedly decreased during the second winter to 4.0 g m−2. The maximum sediment yield, 3436.8 g m−2, was generated in the first winter during a 126 mm rainstorm. To enable comparison among seasons and controlling factors, plots and rain events, we normalized the data per m2 and per mm of precipitation, generating a
244
L. Wittenberg et al. / Catena 121 (2014) 241–247
Fig. 2. Vegetation cover regeneration — north and south aspects, and mean vegetation cover during the 24 month study period.
sediment yield coefficient, which is essentially similar to the RC index. Similarly to the runoff, sediment yields during both rain seasons were significantly higher on the southern aspects compared to the northern ones (Table 3). Likewise, soil losses were significantly higher on steep slopes compared to moderate ones. To compare within-year differences, in response to the different controlling factors, the dataset was partitioned by winter season and each of the sub-sets was analyzed for each of the controlling factors. Considering the problems associated with multiple comparisons we used a conservative rejection level of α = 0.05 / 3 = 0.0166 to detect significant differences. Mean normalized soil losses exhibited a decreasing trend of an order of magnitude between the first (x = 0.71 g m−2 mm−1) and the second (x = 0.08 g m−2 mm−1) winter seasons. The strong decrease in sediment yield across the years and with respect to the different controlling factors was significantly different in all the categories compared. The highest decrease in soil loss amounts (per unit area, per mm of rainfall) between the first and second seasons occurred at the southern aspects and on the steep-slope plots, where absolute values decreased from 20.9 to 6.2 g m−2 mm−1 and from 16.9 to 5.4 g m−2 mm−1, respectively. In relative terms (i.e. proportional decrease), however, the most pronounced reduction occurred in the moderate fire-severity plots, where inter-season values differed by a factor of 25, representing a decrease from 1.78 to 0.19 g m−2 mm−1 (Table 3). The associated standard deviations also decreased between the first and second wet seasons, indicating lower fluctuations in soil loss amounts between events and plots during the second season (see Table 3).
4.3. Regression analyses — runoff and sediment yield Multiple regression analyses were applied to each of the controlling factors in order to analyze the soil loss mechanisms and identify the Table 2 Mean RC and standard deviation (sd) of differing physiographic attributes during the first and second winters. Δt significance denotes differences between the two seasons. Δ factor significance indicates differences between two counterpart sets of plots. 2005–2006
North aspect South aspect Δ factor significance Moderate slope Steep slope Δ factor significance Mod fire severity High fire severity Δ factor significance
2006–2007 sd
x RC (%)
sd
7.8 23.7 0.0008 11.8 16.6 0.0017 10.8 17.9 0.051
9.7 25.6
5.6 17.1 b0.0001 6.01 13.6 b0.0001 8.7 11.8 0.12
8.8 18.0
0.15 0.40
16.3 10.2
0.40 0.47
10.7 17.4
0.847 0.13
15.4 22.5
Table 3 Mean soil losses (g m−2 mm−1) and standard deviation (sd) on differing physiographic attributes during the first and second wet winters. Δt significance denotes differences between the two seasons. Δ factor significance indicates differences between two counterpart sets of plots. 2005–2006
Δt significance (p-val)
x RC (%)
19.5 19.7
significant parameters associated with this process (Table 4). Consequently, the role of five independent variables was assessed during the regression analyses: precipitation, runoff, precipitation days, 30-min storm intensity and vegetation cover. Precipitation reflects the event cumulative rainfall amounts. Runoff is the amount of running water collected following each rain event. ‘Precipitation days’ is defined as the amount of precipitation in a rain event previous to the measured one, divided by the number of dry days between the successive events, i.e., the number of dry days between the previous and measured runoff event (precip-days = mme − 1 / dry days(e − 1,e)). This index served as a proxy for soil moisture, which was included as an additional independent variable in the analysis. The 30-min storm intensity was based on the maximum measured rainfall amounts per 30 min during each event. This variable, however, was never found to be significant. The adjusted R2 of the regression models explained between 7 and 32.1% of the variation, and was highest in the second-winter model. The models suggest that the major mechanism controlling soil loss is the runoff, which appeared to be a significant factor in all the regression models. Vegetation cover and precipitation were significant factors in six of the nine analyses performed, while “precip-days” was a significant variable only in three models. The analysis of the entire dataset yielded three significant variables driving sediment yield: runoff, vegetation cover and precipitation. Comparing between the first and second winter seasons (Table 4) suggests that while runoff was a significant factor during both winters, its relative contribution to sediment yield was seven times higher during the first winter (β = 3.58, β = 0.5 second winter, β's being the parameter coefficients in the regression models). Further, during the first winter the precipitation amounts per se was a significant factor driving sediment
−2
x (g m North aspect South aspect Δ factor sig. Man–W-U Moderate slope Steep slope Δ factor sig. High fire severity Moderate fire Δ factor sig.
Δt sig.
2006–2007 −1
mm
)
sd
−2
x (g m
−1
mm
)
sd
0.6 20.9 0.006
1.55 5.74
0.06 2.2 b0.001
0.13 6.21
b0.001 0.03
0.45 16.9 b0.001 16.4 1.78 b0.001
1.33 51.8
0.05 1.7 b0.001 1.8 0.07 0.005
0.11 5.4
0.01 b0.001
5.7 0.19
0.001 0.01
51.9 11.2
L. Wittenberg et al. / Catena 121 (2014) 241–247
245
Table 4 Soil loss regression model components — β coefficient values.
1 2 3 4 5 6 7 8 9
Controlling factor
Runoff (L mm−1)
Vegetation cover (%)
Event precipitation
Precip.-days
Intercept
Adj. R2
df
Overall model North aspect South aspect Steep slope Moderate slope High fire severity Moderate fire First winter Second winter
1.55 2.28 2.27 21.5 1.28 21.86 10.1 3.58 0.50
−0.013 −0.0009 −0.033 −0.36 −0.009 −0.23 – – –
0.001 – 0.024 0.199 0.005 0.18 – 0.017 –
– – – – 0.09 – 0.51 – −0.005
0.05 0.04 −0.13 12.4 −0.37 −0.8 −4.46 −0.77 0.22
0.15 0.10 0.20 0.22 0.16 0.2 0.07 0.287 0.32
255 136 107 139 101 128 116 99 144
yield, while during the second one it was more closely associated with precipitation amounts. A comparison between the north and south aspects indicates that at the southern aspects the magnitude of the effect of runoff is almost similar to the northern ones. In contrast, the effect of runoff on sediment yield is almost 17 times higher on steep slopes (β = 21.5) compared to moderate slopes (β = 1.28), and twice as high at the severely burnt (β = 21.86) plots compared to the moderately burnt ones (β = 10.1). The models also show vegetation cover to be a dominant parameter negatively correlated with soil loss. A comparison between the coupled controlling factors indicates that the importance of vegetation (or lack of it) is 36 times higher at southern aspects (β = − 0.033) compared to the northern ones (β = −0.0009), and 40 times higher at the steep slopes (β = −0.36) compared to the moderate ones (β = −0.009). Precipitation also appeared to be an important parameter in determining soil loss, and its relative importance was always more prominent in what we consider the extreme conditions, i.e. at the steeper slopes, southern aspects, high severity burnt plots and during the first winter season (Table 4). We also assessed the relationship between runoff and sediment yield, normalized over the amount of precipitation per rain event, during the two rain seasons. During the first season the relationship between these two parameters was weak, as expressed by a lower R2 value of 0.158. During the second season this relationship was stronger, yielding an R2 value of 0.314 (Fig. 3). Each unit of runoff generated less sediment yield during the second season as can be observed by the moderate slope of the regression line during that season (1st year slope = 2.01, 2nd year slope = 0.499). This may suggest that as the relationship becomes stronger between runoff and sediment yield, similar explanatory variables drive both of them.
The vegetation recovery rates, measured in this study, were considerably lower during the first year (26.6%) compared to the second one (62.5% vegetation cover), during which it increased by ~ 36%. By the end of the study period, two years following the fire, vegetation cover had almost reached its pre-fire values. As stated above, it has been suggested that a threshold of 20%–45% coverage is required to considerably reduce runoff yields and soil loss from burnt plots (Inbar et al., 1998; Mayor et al., 2007; Poser and Williams, 1998). Dieckmann et al. (1992) indicated that a 30% vegetation cover reduces the sediment yield by 90% on slopes of less than 30%. Clearly, this threshold has been achieved and even exceeded, but resulted only in a slight, non-significant decrease in runoff yields between the first and second winter seasons. Runoff yields remained relatively high throughout the 2-year study period, even in cases where vegetation cover values were similar to those of the control plot. These findings contradict the commonly reported results which indicate a significant decrease in runoff and sediment yields following the first post-fire winter (Inbar et al., 1998; Puigedefabregas, 2005).
5. Discussion In this study, we established a factorial experimental design to investigate the interconnected effects of slope gradient, aspect and fire severity, and their interrelationships with vegetation recovery patterns, on runoff and erosion processes in a burnt Mediterranean maquis. Runoff and soil loss amounts were commonly significantly higher on the south aspect, on steep gradients and with high fire severity, compared to the north aspect, moderate slopes and moderate fire severity. These results are in accordance with previous fire-related studies conducted in Mediterranean type ecosystems (Campo et al., 2006; Cerda et al., 1995; Gimeno-Garcia et al., 2007; Keizer et al., 2005; Marques and Mora, 1992; Pierson et al., 2002). However, given the experimental design and the marked differences between the burnt plots, particular attention was given to the inconclusive role of plant cover in determining runoff and erosion processes at the particular conditions. Previous studies suggest that post-fire elevated rates of runoff and erosion decline exponentially as vegetation cover increases. A comprehensive review provided by Zuazo and Pleguezuelo (2009) demonstrated that the exponent coefficient values of post-fire elevated rates of runoff and erosion ranges from 0.01 to 0.08, suggesting similar decrease rates in both of these processes. The results of this study, however, do not fall in line with these observations.
Fig. 3. The relationship between runoff and sediment yield during the two winter seasons: a) 2005–6 winter, b) 2006–7 winter. Note the differences in y-axis scale.
246
L. Wittenberg et al. / Catena 121 (2014) 241–247
Presumably, following forest fires, vegetation cover alone is not sufficient to control post-fire hydrological and sedimentological processes and additional properties of the vegetation, such as type and architecture (Zuazo et al., 2004), as well as physio-chemical changes of the soil, play a critical role in determining runoff patterns. As noted in the Results section, RC did not significantly decrease between the two winter seasons, whereas within each year significant differences were observed between the north and south aspects, and the moderate and steep slopes. Variations in fire severity, however, were not sufficient to generate significant differences in runoff. These results suggest that in the case of the Carmel 2005 fire vegetation and topography were key factors in determining post-fire runoff patterns, compared to variation in fire severity which did not generate such a response. In the light of the above discrepancies in runoff and erosion rates and to better understand the specific role of the intra- and inter-season soil yield generating mechanisms, multiple regression analyses were conducted (Table 4). The most influential parameter, which was positively correlated with soil loss, was the runoff. However, runoff did not significantly decrease between the first and second winter seasons, thus it cannot be held accountable for the inter-season decrease in soil loss. We assume that runoff, which depends on the coupled effects of event-specific precipitation properties and soil–vegetation condition, determines soil loss trends which occurred during each of the two rainy seasons. Vegetation cover also emerged as a significant factor driving sediment yield. Regression results point to the differential importance of the vegetation in the “extreme” conditions (steep slopes, severe fires and southern aspect) compared to the moderate ones. In the extreme conditions the absolute values of the vegetation β coefficients were much higher. This suggests that any additional unit of vegetation cover results in larger reductions of sediment yield. Thus we suggest that in “extreme” landscapes geomorphic sensitivity to fire disturbance is moderated to a larger extent by vegetation recovery rates and properties. While soil loss exhibited a decreasing trend during the study period, runoff did not display a similar response. Possibly, these patterns might be explained by the differential effect that vegetation has on these processes. For example, during the first winter season a RC of 2.1% produced 20.5 g m−2, while during the second winter a higher RC (3.6%) produced only 5.1 g m−2. This might indicate that vegetation cover values are sufficiently high to moderate soil loss processes. These responses should also be interpreted in light of the results presented in Fig. 3, which indicate that during the second winter runoff produces less soil loss, as expressed by the slope of the lines, but a higher proportion of the variation in soil loss is accounted for by runoff. Thus a complex set of interactions emerged in this system: runoff did not respond to changes in vegetation cover, whereas soil loss did, and at the same time the correlation between runoff and soil loss increased. This may suggest that other factors which affected runoff and soil loss during the first winter season became less dominant during the second one. Such factors may include presence of ash and hydrophobicity which have been observed to have relatively short term effect Bodí et al. (2012). Alternatively, soil resources could have been exhausted during the first winter, resulting in lower loss values during the following winter. Soil loss may be directly affected by canopy regeneration, whereas runoff is hypothesized to be determined by plant architecture, such as plant canopy and stem structure (Domingo et al., 1998; Martinez-Meza and Whitford, 1996; Zuazo et al., 2004), and possibly the spatial pattern of the vegetation (Mayor et al., 2008). Therefore, a sorting pattern emerges, in which soil loss dynamics change at a certain rate in response to the disturbance event, while runoff yields respond at a different rate. This allows addressing two central aspects; the first, a scientific geomorphological approach, is aimed to better understand runoff and soil loss processes, in the light of vegetation recovery dynamics, while the other is an applied approach, aiming to locate vulnerable areas during the rehabilitation process, in order to minimize the damage inflicted by wildfires.
6. Conclusions The resilience of an eco-geomorphic system, namely, its ability to absorb a perturbation and to recover to its pre-disturbance structure and state, is a vital intrinsic property of the system (Westman, 1986). Mediterranean ecosystems are largely considered as being resilient to fire, soil disturbances and grazing (Lavorel, 1999). The resilience to fire, however, may fundamentally differ within the disturbed area and among the various patches of the burnt landscape, given the specific biotic and physiographic interactions occurring within each of the areas. Essentially, this study was concerned with the geomorphic response of perturbed geo-systems characterized by different physiographic conditions. Runoff and sediment yield from the burnt plots responded differentially to the increase in post-fire vegetation cover, where runoff response seems to lag behind soil loss. Two years after the fire vegetation cover was almost completely restored in terms of areal cover, although not necessarily with respect to its structure. Nonetheless, we suggest that the mere increase in vegetation cover is not sufficient to significantly affect reduction in runoff, and additional properties of the vegetation ought to be considered, assumingly plant architecture and its spatial distribution. Further, the results indicate that the increase in vegetation cover has a more pronounced effect in locations characterized by higher geomorphic sensitivity (sensu Swanson, 1981).
References Bodí, M.B., Doerr, S.H., Cerdà, A., Mataix-Solera, J., 2012. Hydrological effects of a layer of vegetation ash on underlying wettable and water repellent soil. Geoderma 191, 14–23. Campo, J., Andreu, V., Gimeno-Garcia, E., Gonzalez, O., Rubio, J., 2006. Occurrence of soil erosion after repeated experimental fires in a Mediterranean environment. Geomorphology 82, 376–387. Carrión, J.S., Sanchez-Gomez, P., Mota, J.F., Yll, R., Chain, C., 2003. Holocene vegetation dynamics, fire and grazing in the Sierra de Gador, southern Spain. The Holocene 13, 839–849. Cerda, A., Imeson, A., Calvo, A., 1995. Fire and aspect induced differences on the erodibility and hydrology of soils a La Costera, Valencia, Southern Spain. Catena 24, 289–304. Certini, G., 2005. Effects of fire on properties of forest soils: a review. Oecologia 143, 1–10. Cocke, A.E., Fulé, P.Z., Crouse, J.E., 2005. Comparison of burn severity assessments using Differenced Normalized Burn Ratio and ground data. Int. J. Wildland Fire 14, 189–198. Daly, E., Porporato, A., Rodrigez-Iturbe, I., 2004. Coupled dynamics of photosynthesis, transpiration and soil water balance I. Upscaling from hourly to daily time scale. J. Hydrometeorol. 5, 546–558. De Luis, M., Garcia-Cano, M.F., Cortina, J., Raventos, J., Gonzalez-Hidalgo, C.J., Sa'nchez, J.R., 2001. Climatic trends, disturbances and short-term vegetation dynamics in a Mediterranean shrubland. For. Ecol. Manag. 147, 25–37. DeBano, L.F., 2000. Water repellency in soils: a historical overview. J. Hydrol. 231–232, 4–32. Dieckmann, H., Motzer, H., Harres, H.P., Seuffert, O., 1992. Vegetation and erosion. Investigation on erosion plots in Southern Sardinia. Geooko-plus 3, 139–149. Doerr, S.H., Shakesby, R.S., Walsh, R.P.D., 2000. Soil water repellency: its characteristics, causes and hydro-geomorphological consequences. Earth-Sci. Rev. 51, 33–65. Doerr, S.H., Woods, S.W., Martin, D.A., Casimiro, M., 2009. Natural background soil water repellency in conifer forests of the north-western USA: its prediction and relationship to wildfire occurrence. J. Hydrol. 371, 12–21. Domingo, F., Sanchez, G., Moro, M., Brenner, A., Puigdefabregas, J., 1998. Measurement and modeling of rainfall interception by three semi-arid canopies. Agric. Meteorol. 91, 272–292. ESRI, 2009. ArcGIS. Environmental Systems Resource Institute, Redlands, CA, USA. Gimeno-Garcia, E., Andreu, V., Rubio, J., 2007. Influence of vegetation recovery on water erosion at short and medium-term after experimental fires in a Mediterranean shrubland. Catena 69, 150–160. González-Pérez, J.A., González-Vila, F.J., Almendros, G., Knicker, H., 2004. The effect of fire on soil organic matter — a review. Environ. Int. 30 (6), 855–870. Halfon, N., 2003. Translocation of the orographic rainfall enhancement from the ridge towards the lee slopes in northern Israel, unpublished M.Sc. thesis. Haifa University, Israel. p. 145 (in Hebrew). Inbar, M., Tamir, M., Wittenberg, L., 1998. Runoff and erosion processes after a forest fire in Mount Carmel, a Mediterranean area. Geomorphology 24, 17–33. Keizer, J.J., Coelho, C.O.A., Shakesby, R.A., Domingues, C.S.P., Malvar, M.C., Perez, I.M.B., Matias, M.J.S., Ferreira, A.J.D., 2005. The role of soil water repellency in overland flow generation in pine and eucalypt forest stands in coastal Portugal. Aust. J. Soil Res. 43, 337–349. Kutiel, P., Naveh, Z., 1987. The effect of fire on nutrients in a pine forest soil. Plant Soil 104, 269–274. Lavorel, S., 1999. Ecological diversity and resilience of Mediterranean vegetation to disturbance. Diversity and Distributions 5 (1–2), 3–13.
L. Wittenberg et al. / Catena 121 (2014) 241–247 Loch, R., 2000. Effects of vegetation cover on runoff and erosion under simulated rain and overland flow in a rehabilitated site on the Meandu Mine, Tarong, Queensland. Aust. J. Soil Res. 38, 299–312. Malkinson, D., Wittenberg, L., 2007. Scaling the effects of riparian vegetation on crosssectional characteristics of ephemeral mountain streams — a case study of Nahal Oren, Mt. Carmel, Israel. Catena 69, 103–110. Marques, M., Mora, E., 1992. The influence of aspect on runoff and soil loss in a Mediterranean burnt forest (Spain). Catena 19, 333–344. Martinez-Meza, E., Whitford, W., 1996. Stemflow, throughfall and channelization of stemflow by roots in three Chihuahuan desert shrubs. J. Arid Environ. 32, 271–287. Mataix-Solera, J., Cerdà, A., Arcenegui, V., Jordán, A., Zavala, L.M., 2011. Fire effects on soil aggregation: a review. Earth Sci. Rev. 109, 44–60. Mayor, A.G., Bautista, S., Llovet, J., Bellot, J., 2007. Post-fire hydrological and erosional responses of a Mediterranean landscape: seven years of catchment-scale dynamic. Catena 71, 68–75. Mayor, A.G., Bautista, S., Small, E.E., Dixon, M., Bellot, J., 2008. Measurement of the connectivity of runoff source areas as determined by vegetation pattern and topography: a tool for assessing potential water and soil losses in drylands. Water Resour. Res. 44, w10423. Neary, D.G., Klopatek, C.C., DeBano, L.F., 1999. Fire effects on belowground sustainability: a review and synthesis. For. Ecol. Manag. 122, 51–71. Neary, D.G., Ryan, K.C., DeBano, L.F., 2005. Wildland Fire in Ecosystems, Effects of Fire on Soil and Water. General Technical Report RMRS-GTR-42 — Volume 4 USDA, Rocky Mountain Research Station (252 pp.). Ne'eman, G., Goubitz, S., Nathan, R., 2004. Reproductive traits of Pinus halepensis in the light of fire — a critical review. Plant Ecol. 171, 69–79. Pausas, J., Verdú, M., 2005. Plant persistence traits in fire-prone ecosystems of the Mediterranean Basin: a phylogenetic approach. Oikos 109, 196–202. Pausas, J.G., Llovet, J., Rodrigo, A., Vallejo, R., 2008. Are wildfires a disaster in the Mediterranean basin? — A review. Int. J. Wildland Fire 17, 713–723. Pierson, F., Carlson, D., Spaeth, K., 2002. Impacts of wildfire on soil hydrological properties of steep sagebrush-steppe rangeland. Int. J. Wildland Fire 11, 145–151. Porporato, A., Do'dorico, P., Laio, F., Rodrigez-Iturbe, I., 2003. Hydrologic controls on soil carbon and nitrogen cycles. I. Modeling scheme. Adv. Res. Water Resour. 26, 45–58.
247
Poser, I.P., Williams, L., 1998. The effect of wildfire on runoff and erosion in native Eucalyptus forest. Hydrol. Process. 12, 251–265. Puigedefabregas, J., 2005. The role of vegetation patterns in structuring runoff and sediment fluxes in dryland. Earth Surf. Process. Landf 30, 133–147. Rulli, M.C., Bozzi, S., Spada, M., Bocchiola, D., Rosso, R., 2006. Rainfall simulations on a fire disturbed Mediterranean area. J. Hydrol. 327, 323–338. 2007SAS Institute Inc., 2009. JMP. v. 9. Cary, North Carolina, USA. Shakesby, R.A., 2011. Post-wildfire soil erosion in the Mediterranean: review and future research directions. Earth Sci. Rev. 105, 71–100. Soil Survey Staff, 2006. Keys to Soil Taxonomy, 10th ed. NRCS, Washington, DC. Swanson, F.J., 1981. Fire and geomorphic processes. In: Mooney, H.A., Bonnicksen, T.M., Christiansen, N.L., Lotan, J.E., Reiners, W.A. (Eds.), Fire Regime and Ecosystem Properties, United States Department of Agriculture, Forest Service, General Technical Report WO. United States Government Planning Office, Washington, DC, 26, pp. 401–421. Tessler, N., Wittenberg, L., Malkinson, D., Greenbaum, N., 2008. Short term changes in soil water repellency following a forest fire in Mount Carmel, Israel. Catena 74, 185–191. Valentin, C.D., d'Herbes, J.M., 1999. Niger bush as a natural water harvesting system. Catena 37, 1–24. Westman, W.E., 1986. Resilience: concepts and measures. in Resilience in mediterraneantype ecosystems (ed. by B. Dell, A.J.M. Hopkins & B.B. Lamont), pp: 5–19. Wittenberg, L., 2012. Post-fire soil ecology: properties and erosion dynamics. Isr. J. Ecol. Evol. 58 (2), 151–164. Wittenberg, L., Inbar, M., 2009. The role of fire disturbance on runoff and erosion processes — a long-term approach, Mt. Carmel case study, Israel. Geogr. Res. 47, 46–56. Woods, S.W., Balfour, V.N., 2010. The effects of soil texture and ash thickness on the postfire hydrological response from ash-covered soils. J. Hydrol. 3933, 274–286. Zuazo, V.D., Pleguezuelo, C.R.R., 2009. Soil-Erosion and Runoff Prevention by Plant Covers: A Review. Sustainable Agriculture. Springer, Netherlands, pp. 785–811. Zuazo, V.D., Martinez, J.F., Raya, A., 2004. Impact of vegetation cover on runoff and soil erosion at hillslope scale in Lanjaron, Spain. Environmentalist 24, 39–48.