Catena 108 (2013) 50–57

Contents lists available at SciVerse ScienceDirect

Catena

j ourna l homepage: www.e lsev ie r .com/ locate /catena

Effect of forest floor characteristics on water repellency, infiltration, runoff and soilloss in Andisols of Tenerife (Canary Islands, Spain)

J. Neris ⁎, M. Tejedor, M. Rodríguez, J. Fuentes, C. JiménezDpto. Edafología y Geología, Facultad de Biología, Universidad de La Laguna, Av. Astrofísico Fco. Sánchez s/n, 38071, La Laguna, Tenerife, Canary Islands, Spain

⁎ Corresponding author. Tel.: +34 922318366; fax: +E-mail address: jneris@ull.es (J. Neris).

0341-8162/$ – see front matter © 2012 Elsevier B.V. Alldoi:10.1016/j.catena.2012.04.011

a b s t r a c t

a r t i c l e i n f o

Keywords:

AndisolsForest floorWater repellencySoil structureInfiltration rateRainfall simulation

Given its singular properties and location, forest floor (litter+duff) is a key factor in hydrological processes.Water infiltration research was carried out for the present study in Andisols at ten sites, six of which had cov-erings of pine forest and four of rainforest. Rainfall simulations were conducted on gentle, moderately-steepand steep slopes (10, 30 and 50%) to determine infiltration, runoff and soil loss as a function of the forest floorcharacteristics. The duff on the pine forest soils consists of moderately porous, extremely hydrophobic andconsistent semi-decomposed organic material, which is rich in fungi hyphae. The duff on the rainforestsoils is formed by highly porous, loose, semi-decomposed organic material. The study results highlight theinfluential role played by the forest floor in infiltration and runoff. Infiltration barely reaches 20 mm h−1

in pine forest, compared to 50 mm h−1 in rainforest. As a consequence, the pine forest runoff is twice thatrecorded in rainforest sites. The wetting front on gentle and moderately-steep slopes evidences the influenceof the duff on infiltration. In pine forest, most of the rainwater remains in the duff and infiltration dependslittle therefore on the underlying mineral soil properties. In rainforest, the wetting front extends below theduff and the well-developed soil structure is a major factor in water infiltration. The differences noted inthe two parameters are not found on the steep slopes. No soil loss differences are observed between thetwo vegetation covers and forest floors despite the greater runoff in pine forest. The results demonstratethe protective effect of the organic covering and how the stability of the Andisols helps combat water erosionprocesses.

© 2012 Elsevier B.V. All rights reserved.

1. Introduction

As the interface between the mineral soil and the atmosphere inmany land ecosystems, the forest floor – a surface horizon mainly com-prising decomposing plantmaterial – plays a crucial role in hydrologicalprocesses (Keith et al., 2010a). Forest floor is divided into two differen-tiated layers (litter and duff) following a gradient of decomposition(Keith et al., 2010b). As the above authors note, litter (L or A00 horizon)is formed by fresh leaves from the surrounding vegetation and is lo-cated on the surface. Duff (A0) comprises two distinct layers: a top fer-mentation layer (F horizon) and bottom humus layer (H horizon). Thelayers are formed, respectively, by partly or wholly decomposed plantremains.

Several authors have drawn attention to the physical particulari-ties of the forest floor and note how they differ to those of the under-lying mineral soils (Keith et al., 2010b; Lauren and Mannerkoski,2001; Lauren et al., 2000). Due to its location and properties, the pres-ence of forest floor can modify the amount of rainwater available forinfiltration and runoff (Guevara-Escobar et al., 2007). Consequently, it

34 922318311.

rights reserved.

can alter the hydrological response of the soils compared to theirresponse when no forest floor is present (Keith et al., 2010a). Accord-ing to Descroix et al. (2001), surface features play a pivotal role in soilhydrology. However, most hydrological studies in forest zones focuson the mineral soil and few take the forest floor into consideration(Buttle et al., 2000, 2005). Furthermore, it is well known that thepresence of water-repellent soil surface horizons inhibits infiltrationand promotes runoff (DeBano, 1971; Doerr et al., 2000; Robichaudand Waldrop, 1994). However, most studies concern themselveswith the underlying mineral horizons and few with the forest floor(see e.g. Martínez-Zavala and Jordán-López, 2009; Poulenard et al.,2001; Zehetner and Miller, 2006).

Andisols (Soil Survey Staff, 1999) are the most characteristic soilsfound on Tenerife (Canary Islands, Spain) due to the island's volcanicnature, the relatively young age of some of its materials and its appro-priate climate conditions. These are soils with high structural devel-opment and high porosity, properties which explain their high waterinfiltration rate under natural conditions (Harden, 1991; Nanzyoet al., 1993; Perrin et al., 2001). Most authors consider that the singu-lar mineralogical properties and high organic carbon content of thesesoils are decisive factors in their structural properties (FernándezCaldas and Tejedor Salguero, 1975; Hoyos and Comerford, 2005;Quantin, 1994; Warkentin and Maeda, 1980). Nonetheless, previous

51J. Neris et al. / Catena 108 (2013) 50–57

field observations show that, in pine forest, some of the soils donot become moist until the end of the rainy season. This wettingdelay would appear to be in contradiction with the theoreticallyhigh infiltration of Andisols. Studies of Andisols have highlighted thestructural degradation and reduction in infiltration capacity producedby a change in land use (see e.g. Jiménez et al., 2006; Poulenard et al.,2001; Rodríguez Rodríguez et al., 2002; Warkentin and Maeda, 1980;Zehetner and Miller, 2006). Meanwhile, various authors have drawnattention to the influence exerted on soil hydrological behaviour bysurface features, including plant cover (Cerdà, 1998, 1999; Molinaet al., 2007), rock fragments (Descroix et al., 2001; Martínez-Zavalaand Jordán, 2008) and even ash (Cerdà and Doerr, 2008; Woods andBalfour, 2008, 2010; Zavala et al., 2009) and pine needles following afire (Cerdà and Doerr, 2008). However, few studies have examinedthe influence of the forest floor on the water repellency and hydro-logical behaviour of Andisols.

The present work aims to contribute to resolving the discrepanciesthat exist with respect to thewetting process in certain Andisols underforest vegetation. In order to do so we will: (a) characterise the forestfloor and soil for two different vegetation covers; (b) study the infil-tration, water repellency, runoff and erosion in the Andisols underboth covers using simulated rainfall events; and (c) analyse the influ-ence of the forest floor and soil properties on the hydrological process-es in the study soils.

2. Methodology

2.1. Site description

Tenerife (Canary Islands, Spain) is an island in the Atlantic Ocean.It is situated between 27º 55′ and 28° 35′ north latitude and between16° 05′ and 16° 55′ west longitude (Fig. 1). Its geographical position(near the Tropic of Cancer and under the influence of the tradewinds), elevation (highest point: 3718 m.a.s.l.) and the orientationof its mountain systems give rise to a wide variety of meso- andmicro-climates and vegetation (del Arco et al., 2006). It also boastsa diversity of volcanic materials of different ages. The combinationof all these factors accounts for the presence of different soil orderson the island (Tejedor et al., 2007).

The study area is situated between 850 and 1400 m.a.s.l. on thenorth face of the island (Fig. 1). It is dominated by moderately-steephillsides (20–50%) according to the slope classification given in the

Fig. 1. Location of the island of Tenerife; Andis

Soil Survey Manual (Soil Survey Division Staff, 1993). Bedrock con-sists of basaltic pyroclasts and lava flows (0.7–0.01 M yrs) with subse-quent rejuvenations by analogous ashes (b0.01 M years). Averageannual precipitation is between 600 and 1000 mm. The amount ofwater from condensation is very significant in this altitudinal strip.According toMarzol Jaén (2005), water from condensation can amountto five times rainfall, depending on the vegetation and location. Thenatural vegetation consists mainly of pine forest (Pinus canariensis)and rainforest (Laurus novocanariensis, Apollonias barbujana, Perseaindica, Ilex canariensis, Myrica faya, Erica arborea, and Erica scoparia,among other species).

The study area lies on the ustic–udic boundary (Soil Survey Staff,1999). The soils are mostly Andisols (Ustands and Udands). SomeInceptisols and Entisols (Soil Survey Staff, 1999) are also found al-though they occupymuch smaller areas. The type of vegetation deter-mines the characteristics of the forest floor, particularly the duffproperties. The litter in both the pine forest and rainforest consistsof a covering of loose leaves in the early stages of decomposition.The pine forest duff comprises semi-decomposed organic material,which is rich in macroscopic fungi hyphae (Fig. 2). In contrast, therainforest duff is a loose, semi-decomposed organic material whichappears to be free of fungi hyphae (Fig. 2).

2.2. Site selection

Six pine forest sites and four rainforest sites were selected for thestudy. All the soils are Udands or Ustands (Soil Survey Staff, 1999).The sites are located at heights between 900 and 1200 m.a.s.l. Bearingin mind the influence of vegetation cover on forest floor characteris-tics, care was taken to select sites with homogeneous vegetationcomposition, both in the pine forest and rainforest. Pinus canariensisis the main species present in the pine forest sites, while the vegeta-tion in the rainforest sites comprises laurifolia tree species (Myricafaya and Persea indica). The limited extent of rainforest on the island,its location in areas that are difficult to access and its wide botanicalcomposition prevented further rainforest sites from being selected.3 plots with similar forest floor characteristics were selected in eachsite for each of the study slopes: 10% (30 plots), 30% (30 plots) and50% (15 plots). Only 5 sites with slopes of 50% were identified (2 inpine forest and 3 in rainforest). In line with the Soil Survey DivisionStaff (1993) slope classes, the terms gentle, moderately-steep andsteep were used for slopes of 10, 30 and 50%, respectively. The

ols (Udands and Ustands) and study sites.

Fig. 2. Above, left to right — forest floor, duff specimen and water repellency in pine forest. Below, left to right — forest floor, litter and duff samples in rainforest.

52 J. Neris et al. / Catena 108 (2013) 50–57

research (site description, forest floor characterisation, sample collec-tion and simulated rainfall events) was conducted from July toSeptember 2007, following a dry period (total rainfall during the pre-vious month was below 10 mm).

2.3. Forest floor characterisation

A full description of the forest floor at each site was undertakenduring the summer of 2007 prior to the simulated rainfall events.Litter and duff were moderately dry in all sites. The thickness andcovering of both the litter and duff were studied. Duff was also char-acterised in terms of consistency, presence of macroscopic fungi andwater repellency. The thickness and covering of the litter and duffwere estimated visually in each study plot. A stereoscopic microscopewas used to check the presence of fungi hyphae. Duff consistencywas determined by hand in moderately-dry samples using ruptureresistance classes (Soil Survey Division Staff, 1993). The Water DropPenetration Time test (WDPT) (Letey, 1969) was used to determinewater repellency. The test consists of placing 10 water drops on thesample surface and measuring the time taken to infiltrate completely.Five air-dried sub-samples from each site were analysed using thismethod. When WDPT exceeded 1 h samples were covered carefullyto prevent evaporation.

2.4. Infiltration, runoff and soil loss

A rainfall simulator similar to that developed by Nacci and Pla(1991) was used to study infiltration, runoff and soil loss. The gravitydrip simulator was placed at a height of ≈2 m over a 0.1 m2 rectan-gular bounded parcel. Parcel boundaries were established using a0.27×0.37×0.25 m frame nailed 0.05 m into the ground. Plasticsheeting was wrapped around the simulator to minimise rainfall var-iability due to wind. Bearing in mind the high infiltration capacity ofthe Andisols, high rainfall intensity (60 mm h−1) was used to guaran-tee runoff phenomena. High intensity rainfall events were simulatedto quantify the effects of intense thunderstorms, which are the onesthat trigger erosive processes. Rainfall intensity was calculated beforeand after each event by measuring the volume of water collected dur-ing a 5-min interval in a plastic calibration pan placed over the plotframe. Instant rainfall intensity was determined by linear regressionbetween the intensity before and after the experiment. Rainfall

events were of 35 min duration. Demineralised water was usedgiven that electrolyte concentration may affect rainfall simulationresults (Agassi and Bradford, 1999; Borselli et al., 2001). A total of75 rainfall simulations were performed.

The rainfall simulations were carried out during the summer of2007 after a dry period. The forest floor was moderately dry and thetopsoil moisture of the study plots was lower than water retentionat permanent wilting point (20±8% and 30±11% in the pine andrainforest sites respectively). The Volume to Runoff (VR: mm) – theamount of rainwater needed to generate runoff – was calculated. Allthe runoff water was measured at 5 minute intervals and collectedat alternate intervals in 500 ml jars for sediment analysis. The Infiltra-tion Rate (IR: mm h−1) until the steady-state was reached (Steady-state Infiltration Rate-SIR: mm h−1) was determined. The Runoff/Rainfall Ratio (RR: %) was calculated as the relationship betweenthe total runoff volume and the volume of rainfall during the rainfallevent. Runoff samples were dried (105 °C) until all the water hadevaporated completely and were then weighed to determine themass of eroded soil. Soil loss was represented as the soil loss rate(SR: g m−2 h−1) and concentration (SC: g L−1). Following the rain-fall event, the wetting front (WF) was determined visually by diggingand was measured from the forest floor surface.

2.5. Soil analysis

Three bulk mineral soil samples were collected from the upper5 cm below duff at each site (A horizon). Particle size distribution(particles b2 mm) was determined after prior H2O2 treatment andsamples were dispersed in sodium hexametaphosphate solutionusing the Bouyoucos densimeter method (Gee and Bauder, 1986).Soil organic carbon (SOC) was determined by dichromate oxidation(Walkley and Black, 1934). Bulk density was measured on an oven-dried weight basis of a 96.6 cm3 core sample taken at field-moistureconditions (Blake and Hartge, 1986). Gravimetric soil moisture at−33 kPa and −1500 kPa tensions was determined by Richards pres-sure plates (Klute, 1986). Three indices associated with soil structurewere studied following the method of Bartoli et al. (1991): soil aggre-gation (SA), wet soil stability (WSS) and wet aggregate stability(WAS). The water repellency of the top 5 cm of soil below duff wasdetermined using the Water Drop Penetration Time (WDPT) test(Letey, 1969). For a description of the method, see Section 2.3 above.

Fig. 3. Influence of vegetation on particle size distribution, organic carbon (OC), soil ag-gregation (SA), wet soil stability (WSS) and wet aggregate stability (WAS), soil waterretention at −33 kPa and −1500 kPa (SWR), bulk density (BD) and logWDPT (Boxplot: thick bar=median; upper and lower limits of the 75 and 25 percentiles, respec-tively; same letter indicates no statistically significant differences between vegetationtype with Mann Whitney U test (pb0.05).

53J. Neris et al. / Catena 108 (2013) 50–57

2.6. Statistical analysis

SPSS version 17.0.0 was used for the statistical analysis of theresults. Given that the data did not satisfy the requirements for nor-mality and homogeneity of variances, non-parametric methodswere used for the analyses. The Mann–Whitney U test, which estab-lishes differences between groups defined for independent samples,was used for hypothesis testing. A significance level of 0.05 was set.

3. Results

3.1. Forest floor characterisation

Table 1 shows the main characteristics of the forest floor of thestudy soils. As can be seen, the covering and thickness of the duffand litter are independent of the type of vegetation cover present.In the pine forest duff consistency is moderately hard in a moderatelydry sample. The duff forms a continuous structure on the soil surfaceat these sites (Fig. 2). In the rainforest the duff covers a large percent-age of the soil surface but is classified as loose in a moderately drysample (Fig. 2). This organic material tends to be rich in macroscopicfungi hyphae (90% of plots) where pine forest is the dominant vege-tation. Conversely, the presence of macroscopic fungi hyphae ismarginal in rainforest and it is found in only 9% of the study plots(Fig. 2). Although all the duffs exhibit high water repellency, statisti-cally significant differences are seen according to the vegetation. TheWDPT test results show that the pine forest duff is more repellentand is classified as extremely repellent, whereas the duff in the rain-forest falls within the severely repellent category.

3.2. Soil characteristics

Fig. 3 shows the results for the main infiltration-related soil prop-erties according to the vegetation cover. Differences are found inproperties such as grain size, aggregation and aggregate stability,and water repellency. All the soils are silty loam or loam in texture.However, the rainforest sites have significantly higher clay contentand the pine forest sites significantly more sand. All the soils havehigh aggregation and high structural stability values, although somedifferences are seen according to the vegetation cover (Fig. 3). Theaverage soil aggregation (SA) values of the rainforest soils areabove 75%, compared to around 50% in the pine forest soils. These dif-ferences are statistically significant. The same is true for wet soil sta-bility (WSS). However, wet aggregate stability (WAS) does not differsubstantially between the two types of vegetation. WAS exceeds 80%in both the pine forest and rainforest. This result indicates that thehigherWSS in the rainforest sites is due chiefly to the greater SA as op-posed to greater WAS. Water repellency is classified as extremely re-pellent for the study zones under both vegetation covers. Only 7%and 17% of the observations classify as non-repellent in the pine forestand rainforest, respectively. However, this parameter presents high

Table 1Forest floor characteristics associated with water infiltration.

Litter Duff

Thickness Coverage Thickness Cover

cm % cm %

Pine forestx 1.6 95.8 1.3 93.3σ 0.6 6.6 0.3 12.1Rainforestx 1.4 93.8 1.5 88.8σ 0.5 7.5 0.4 13.1

x: mean; σ: standard deviation; WDPT: Water Drop Penetration Time.

variability (Figs. 3 and 4) and some degree of dependence on vege-tation is also detected (Fig. 3). The pine forest soils have slightlyhigher WDPT values than their rainforest counterparts. However,the values lack statistical significance due to the high variability ofthe results. No major differences according to vegetation are seenin the case of other properties that influence infiltration, such asorganic carbon, short-range order product content, bulk densityand water retention capacity (−33 kPa and −1500 kPa). Typicallyof soils of this type, high organic carbon values are found (12% onaverage). The sum of Alo and ½ Feo on topsoil ranges from 1.1 to 4.6and Alp/Alo from 0.1 to 0.5. These results indicate a predominance ofsilandic material on the soils (IUSSWorking Group, 2006). Bulk densi-ty values are low, barely reaching 0.7 Mg m−3, and are associatedwith the organic matter content and predominance of short-rangeorder products in these soils. Lastly, bearing in mind the clay content,a higher than expected water retention capacity is found, especiallyat low tensions (−33 kPa). These results can be attributed to theincomplete dispersion of the allophanic soils (Armas-Espinel et al.,2003).

age WDPT Fungi hyphae frequency Consistency

s %

3677.5 90 Moderately hard1115.9

1500.0 9 Loose356.7

Fig. 4. Distribution of water repellency classes for the different vegetation covers.

Fig. 6. Wetting front distribution in rainfall events for the different vegetation covers(measured from soil surface).

54 J. Neris et al. / Catena 108 (2013) 50–57

3.3. Effect of vegetation cover and forest floor on infiltration, runoff andsoil loss

Fig. 5 shows the average values, standard deviation and statisticalsignificance for infiltration, runoff and soil loss according to vegeta-tion and slope. Differences according to vegetation are seen in thehydrological parameters studied (VR, SIR, RR, WF, SR and SC) in thecase of the gentle and moderately-steep slopes. In contrast, for steepslopes the results indicate no differences in any of the parameters.

The pine forest soils show a greater tendency for runoff generationthan the rainforest soils on gentle and moderately-steep slopes(Fig. 5). In both slope categories the runoff values in the rainforestsoils are low (below 20%), approximately half those recorded in thepine forest soils. The differences are statistically significant. Differ-ences are also seen in SIR and VR, although they are less pronounced.The rainforest soils exhibit a high infiltration capacity. SIR in rain-forest is 50 and 40 mm h−1 for the gentle and moderately-steep

Fig. 5. Influence of vegetation and slope on volume to runoff (VR), infiltration (SIR),runoff (RR), wetting front (WF), sediment rate (SR) and sediment concentration (SC)(Box plot: thick bar=median; upper and lower limits of the 75 and 25 percentiles, re-spectively; same letter indicates no statistically significant differences between vegeta-tion types (same slope) with Mann Whitney U test (pb0.05); same number indicatesno statistically significant differences between slope angles (same vegetation type)with Mann Whitney U test (pb0.05).

slopes, respectively. By comparison, the values in pine forest arearound 30 mm h−1 lower for both slopes and rarely exceed 20 mmh−1. The differences are statistically significant. VR differs accordingto vegetation only on the gentle slopes, where the values are approx-imately 1 mm higher in the rainforest than in the pine forest. Again,the differences are statistically significant.

The WF of the soils is consistent with the above results. Measuredfrom the litter surface after 35 min of simulated rainfall, the WF isapproximately 7, 6 and 3 cm in rainforest for the gentle, moderately-steep and steep slopes, respectively (Fig. 5). In the pine forest sitesthe WF ranges from 2 to 3 cm on all three slopes. The WF fails toreach the soil surface (A horizon) on the pine forest slopes and wetsonly the forest floor (Fig. 6). In contrast, in the rainforest the WF pen-etrates the soil in all cases on the gentle and moderately-steep slopes,attaining a maximum wetting of 5 cm. The WF differences betweenpine and rainforest are statistically significant for both the gentleand moderately-steep slopes (Fig. 5).

Although overall the study soils present low SR values, slightdifferences are observed depending on the vegetation cover. SR is ap-proximately 5 and 10 g m−2 h−1 for the gentle andmoderately-steepslopes in the pine forest soils. The values are 2–3 g m−2 h−1 higherthan those obtained in the rainforest soils for the same slopes. No sta-tistical significance is observed in these differences for any of thestudy slopes. The SR differences according to vegetation are muchsmaller than those seen for the runoff processes. In addition, the SCvalues are low for all the study soils. No statistically significant differ-ences according to vegetation are found for this parameter and thehigher SR in the pine forest sites is due therefore to the increase inthe volume of runoff water.

3.4. Slope influence on hydrological properties

In general, a decrease is noted in VR, SIR andWFwith increased slopefor both types of vegetation (Fig. 5). Byway of contrast, RR, SR and SC allincrease (Fig. 5). The magnitude of the changes in hydrological proper-ties with increased slope depends on the vegetation present in the zone.

In pine forest sites the change fromgentle tomoderately-steep slopeproduces non-significant variations in VR, SIR, WF and RR. Much largerincreases are seen in SRwith increasing slope, however. The differencesin SR are approximately 4 and 6 g m−2 h−1, respectively, when theslope changes from gentle to moderately-steep and from moderately-steep to steep. The differences noted in SC with increased slope areunimportant. Statistical analysis indicates statistically significant differ-ences only in the case of SR with the change from gentle to steep slope.

The extent of the changes triggered by increased slope in thehydrological parameters is greater in the rainforest sites. The increase

55J. Neris et al. / Catena 108 (2013) 50–57

from gentle to moderately-steep slope leads to slight falls in VR, SIRand WF (0.5 mm, 7 mm h−1 and 1.2 cm, respectively) and slightlyhigher RR (2%). The differences are greater in the case of SR, whichincreases from 3 to 6 g m−2 h−1. None of the aforementioned differ-ences prove to be statistically significant. However, statistical signifi-cance is found in the case of steep slopes. Falls of 1 mm in VR, 34 mmh−1 in SIR and 3.4 cm in WF are observed with the change frommoderately-steep to steep slopes. A more substantial fall in all threeproperties is noted when the results for gentle and steep slopes arecompared (1.5 mm for VR, 41 mm h−1 for SIR and 4.7 cm for WF).Statistical significance is seen between the results for the differentcategories of slope. Major differences in RR and SR are observed withthe transition to steep slopes: RR increases by 41% and SR by 7 g m−2

h−1 in these slopes compared to the moderately-steep slopes. Thesedifferences increase to 43% and 10 g m−2 h−1, respectively, when gen-tle and steep slopes are compared. Statistical analysis indicates that thedifferences are statistically significant when steep slopes are comparedto both gentle and moderately-steep slopes. As in the pine forest sites,SC in rainforest does not differ appreciably with increased slope.

4. Discussion

The infiltration and runoff differences between the pine forest andrainforest Andisols reflect the influence of the forest floor on waterinfiltration processes. The results for the rainforest sites are similar tothose published by other authors for Andisols under forest cover(Perret et al., 1996; Perrin et al., 2001; Poulenard et al., 2001). In con-trast, the pine forest results conflict with the high infiltration tradition-ally attributed toAndisols under forest vegetation (see e.g. Nanzyo et al.,1993; Poulenard et al., 2001; Zehetner and Miller, 2006).

The comparison of mineral soil surface properties evidences thelower structural development and greater water repellency of thepine forest soils (Fig. 3). These differences may account for the differ-ent infiltration behaviour exhibited by the two types of forest. Thedirect relationship between infiltration and structural developmentis well known for all soils, particularly Andisols (Nanzyo et al., 1993;Poulenard et al., 2001; Rodríguez Rodríguez et al., 2002; Warkentinand Maeda, 1980). However, the structural development values ofthe pine forest soils are not low enough to explain the infiltrationresults obtained here. The SIR values recorded in the pine forest sitesare approximately 10–25 mm h−1 lower than those obtained byPoulenard et al. (2001) for Andisols. However, the soil aggregationand water stability values are comparable to those published by saidauthors. Regarding water repellency, many authors have found thatthe presence of water-repellent soil horizons can inhibit infiltrationand promote runoff (DeBano, 1971; Doerr et al., 2000; Robichaudand Waldrop, 1994). On average, the pine forest sites studied hereexhibit greater water repellency than the rainforest sites, althoughthe values recorded are very high in both cases. Less than 20% of rain-forest sites are classified as non-repellent and hence the differencesbetween the two do not explain their behaviour.

The analysis of the wetting front in the pine forest sites (Fig. 6)shows that the rainwater remainsmostly in the duff and barely reachesthe top 1 cm of the underlying mineral soil (A horizon). Under theseconditions the soil properties of the surface mineral horizon will havevery limited impact on the hydrological behaviour of the soils. In con-trast, in the rainforest sites there is considerable water penetration ofthe soil. In these cases the mineral soil properties determine the soil'shydrological behaviour, which is similar to that found by other studies.The results obtained here indicate the major role played by the duff inwater infiltration processes. As numerous authors have noted, the loca-tion and properties of this horizon can alter the soil's hydrologicalbehaviour (Guevara-Escobar et al., 2007; Keith et al., 2010a,b).

The low level of decomposition results in poor cohesion of theshed litter leaves. Consequently, the litter is unlikely to have anymajor influence on infiltration given that the water can run freely

between the leaves and reach the underlying horizons. However,duff characteristics can differ considerably to those of the litter,given the greater degree of decomposition involved (Fig. 2). Themain duff properties influencing infiltration appear to be cohesionand hydrophobicity. Water repellency is high in both types of vegeta-tion studied. Under natural conditions, water repellency is associatedmainly with the organic compounds derived from the activity ordecomposition of plants and organisms (Doerr et al., 2000). Theamount and type of organic matter affect hydrophobicity (Martínez-Zavala et al., 2009). Some genera such as the Pinus or Erica presentin the pine forest and rainforest are listed among the higher plantspecies associated with water repellency (Cerdà and Doerr, 2005,2007; Doerr et al., 2000). Moreover, the presence of fungi hyphae(foundmainly in the pine forest sites) is also a source of hydrophobic-ity (Doerr et al., 2000; Lin et al., 2006). In water repellent soils therainwater tends to remain on the surface of the hydrophobic horizon(duff) (Doerr et al., 2000) and circulates along it aided by the slope(Martínez-Zavala et al., 2009; Wallach and Jortzick, 2008). However,the presence of discontinuities, cracks or paths in the water-repellent surface provides preferential flow zones for the water(Doerr et al., 2000). In this regard, duff cohesion plays a fundamentalrole. In the rainforest, the poor cohesion of the duff plant remnantshas a positive effect on water infiltration. Despite the duff's high re-pellency, the water is able to circulate through it and reach the under-lying horizons, as the wetting front analysis shows. Consequently, soilproperties such as high aggregation and structural stability are keyfactors in the high infiltration and low runoff values (Mataix-Soleraet al., 2011) noted for these soils on gentle and moderately-steepslopes (Fig. 5). Runoff on steep slopes is greater due primarily to theshorter time the water remains on the soil because of the gravityeffect and reduced effectiveness of the surface roughness and waterstorage capacity (Sharma et al., 1983). In pine forest, the greater co-hesion of the duff plant material results in fewer paths and thereforehinders the water's efforts to reach deeper horizons. Under these con-ditions infiltration is determined fundamentally by the forest floorproperties and less so by the mineral soil, which would explain thelow infiltration and high runoff values observed in these zones. More-over, the limited water infiltration in the pine forest duff would alsoaccount for the negligible infiltration and runoff differences foundwith increased slope. Infiltration is very low and runoff is very highon gentle slopes. Consequently, an increase in slope produces nochanges of note (Fig. 5).

The limited water infiltration in the pine forest soils helps explainprevious field observations indicating a delay in the soil wettingprocess in some zones during the rainy season. It is worth notingalso that the influence of forest floor water repellency on hydrologycan be assumed to be highly seasonal, given the inverse relationshipbetween soil water repellency and soil moisture (DeBano, 1971;Doerr and Thomas, 1998; Huffman et al., 2001; Regalado and Ritter,2009). The gradual wetting of the forest floor and mineral soil surfacehorizons as a result of successive rainfalls reduces water repellencyand therefore promotes water infiltration and wetting processes.The effects of the forest floor on water infiltration and wetting oughtto have a major influence on the moisture regimes of the soils and,consequently, on soil classification. Moreover, the results obtainedhere demonstrate the important role of the forest floor properties inhydrological behaviour. According to Doerr et al. (2000), some of theimpacts of water repellency are enhanced streamflow responses torainstorms and total streamflow. Consequently, the aforementionedproperties should be included as parameters in hydrological modelsused for watershed management. This is particularly appropriate inthe case of Andisols, given their key role in infiltration and runoff.These aspects are the subject of ongoing work.

In contrast to infiltration and runoff, soil loss processes are lessdependent on the type and characteristics of the forest floor. Soilloss is relatively low and the results found here are similar to those

56 J. Neris et al. / Catena 108 (2013) 50–57

reported by other authors for Andisols under natural conditions (Perretet al., 1996; Poulenard et al., 2001; Zehetner and Miller, 2006). Mostauthors note that Andisols are highly resistant to erosion (Nanzyoet al., 1993; Rodríguez Rodríguez et al., 2002; Zehetner and Miller,2006), a circumstance generally attributed to two factors: (i) the soils'high structural stability and low erodibility (Maeda and Soma, 1985;Nanzyo et al., 1993) and (ii) their high porosity and water infiltrationcapacity. The results for structural development (SA, WSS and WAS)and sediment concentration (SC) obtained in the present work are inagreementwith the first of the two premises. The rainforest site findingscoincidewith the observations of the above authors as regards high infil-tration in Andisols. However, the results from the pine forest sites do not.As noted earlier, the characteristics of the forest floor in pine forest act asa constraint on water infiltration. Nonetheless, the sediment rate (SR)differences observed between the two types of vegetation are muchsmaller than those seen for infiltration and runoff. No differences arefound in the case of sediment concentration (SC). The smaller differencesin soil loss processes are attributed to the low erodibility of the soils,which is associated with their structural stability and the protectionafforded by the forest floor against the erosive effects of rain.

Finally, the elimination of the forest floor as a result of a changein land use or management can propitiate extensive erosive phenome-na. It is generally acknowledged that prolonged drying causes irrevers-ible aggregation of the finest grain-size fractions in Andisols (Nanzyoet al., 1993). Consequently, drying triggers changes in their hydricproperties (Poulenard et al., 2004) and increases their hydrophobicity(Podwojewski et al., 2002). These changes impact negatively on theresistance of Andisols to erosion.

5. Conclusions

Forest floor properties such as water repellency are key factors insoil hydrological processes. This organic layer acts as the interfacebetween the soil and atmospheric phenomena, including rainfall. Ingeneral, Andisols have a high infiltration capacity and generate littlerunoff. However, duff characteristics can lead to major changes inthe values of both parameters. As the results obtained here show,the properties of this organic horizon with the greatest hydrologicalimpact are its water repellency, continuity and cohesion.Water repel-lency limits water infiltration while cohesion reduces the presence ofthe preferential flow paths that enable rainwater to reach the under-lying horizons. The wetting front analysis indicates that structural de-velopment retains its influence on infiltration and runoff in rainforest.In pine forest, however, the influence of the soil properties on infil-tration is constrained by the presence of the forest floor. The resultshighlight the need to include forest floor properties, particularly waterrepellency, as parameters in hydrological models. On the other hand,no significant increases in soil loss as a consequence of the lower infil-tration are observed. Erosion values remain low, as many authorshave already noted for the soils of the order. This result demonstratesthe protective effect of the forest floor against erosive processes andthe high resistance of Andisols to erosion under forest conditions.

Acknowledgements

The authors thank Dr. Hernández-Moreno (Universidad de LaLaguna) for reviewing the manuscript. The present research was un-dertaken as part of a project entitled “Characterisation of the soils ofthe island of Tenerife, with particular reference to hydric functions”,funded by the Consejo Insular de Aguas de Tenerife (Tenerife WaterCouncil).

References

Agassi, M., Bradford, J.M., 1999. Methodologies for interrill soil erosion studies. Soil &Tillage Research 49, 277–287.

Armas-Espinel, S., Hernández-Moreno, J.M., Muñoz-Carpena, R., Regalado, C.M., 2003.Physical properties of “sorriba”‐cultivated volcanic soils from Tenerife in relationto andic diagnostic parameters. Geoderma 117, 297–311.

Bartoli, F., Burtin, G., Herbillon, A.J., 1991. Disaggregation and clay dispersion ofOxisols–Na resin, a recommended methodology. Geoderma 49, 301–317.

Blake, G.R., Hartge, K.H., 1986. Bulk density. In: Klute, A.K. (Ed.), Methods of soil analysis.Part I: Physical and mineralogical methods. American Society of Agronomy – SoilScience Society of America, Madison, pp. 363–375.

Borselli, L., Torri, D., Poesen, J., Sanchis, P.S., 2001. Effects of water quality on infiltra-tion, runoff and interrill erosion processes during simulated rainfall. Earth SurfaceProcesses and Landforms 26, 329–342.

Buttle, J.M., Creed, I.F., Pomeroy, J.W., 2000. Advances in Canadian forest hydrology,1995–1998. Hydrological Processes 14, 1551–1578.

Buttle, J.M., Creed, I.F., Moore, R.D., 2005. Advances in Canadian forest hydrology,1999–2003. Hydrological Processes 19, 169–200.

Cerdà, A., 1998. The influence of geomorphological position and vegetation coveron the erosional and hydrological processes on a Mediterranean hillslope. Hydro-logical Processes 12, 661–671.

Cerdà, A., 1999. Parent material and vegetation affect soil erosion in eastern Spain. SoilScience Society of America Journal 63, 362–368.

Cerdà, A., Doerr, S.H., 2005. Influence of vegetation recovery on soil hydrology anderodibility following fire: an 11-year investigation. International Journal of Wild-land Fire 14, 423–437.

Cerdà, A., Doerr, S.H., 2007. Soil wettability, runoff and erodibility of major dry-Mediterranean land use types on calcareous soils. Hydrological Processes 21,2325–2336.

Cerdà, A., Doerr, S.H., 2008. The effect of ash and needle cover on surface runoff anderosion in the immediate post-fire period. Catena 74, 256–263.

DeBano, L.F., 1971. Effect of hydrophobic substances on water movement in soil duringinfiltration. Soil Science Society of America Proceedings 35, 340–343.

del Arco, M., Pérez-de-Paz, P.L., Acebes, J.R., González-Mancebo, J.M., Reyes-Betancort, J.A.,Bermejo, J.A., de Armas, S., González-González, R., 2006. Bioclimatology and climato-philous vegetation of Tenerife (Canary Islands). Annales Botanici Fennici 43, 167–192.

Descroix, L., Viramontes, D., Vauclin, M., Barrios, J.L.G., Esteves, M., 2001. Influence ofsoil surface features and vegetation on runoff and erosion in the western SierraMadre (Durango, northwest Mexico). Catena 43, 115–135.

Doerr, S.H., Thomas, A.D., 1998. The role of soil moisture in controlling water repellency:New evidence from forest soils in Portugal. International Workshop on Soil WaterRepellency, Wageningen, Netherlands, pp. 134–147.

Doerr, S.H., Shakesby, R.A.,Walsh, R.P.D., 2000. Soilwater repellency: its causes, character-istics and hydro-geomorphological significance. Earth-Science Reviews 51, 33–65.

Fernández Caldas, E., Tejedor Salguero, M.L., 1975. Andosoles de las Islas Canarias. Serviciode Publicaciones de la Caja General de Ahorros de Santa Cruz de Tenerife, Santa Cruzde Tenerife. 210 pp.

Gee, G.W., Bauder, J.W., 1986. Particle size-analysis. In: Klute, A. (Ed.), Methods of soilanalysis. Part I. Physical and mineralogical methods. American Society of Agrono-my – Soil Science Society of America, Madison, pp. 383–411.

Guevara-Escobar, A., González-Sosa, E., Ramos-Salinas, M., Hernández-Delgado, G.D.,2007. Experimental analysis of drainage and water storage of litter layers. Hydrologyand Earth System Sciences 11, 1703–1716.

Harden, C.P., 1991. Land-use, soil-erosion, and reservoir sedimentation in an Andeandrainage-basin in Ecuador. International Workshop and Field Excursions: Moun-tain Geoecology of the Andes, Santiago, Chile, pp. 177–184.

Hoyos, N., Comerford, N.B., 2005. Land use and landscape effects on aggregate stabilityand total carbon of Andisols from the Colombian Andes. Geoderma 129, 268–278.

Huffman, E.L., MacDonald, L.H., Stednick, J.D., 2001. Strength and persistence of fire-induced soil hydrophobicity under ponderosa and lodgepole pine, Colorado FrontRange. Hydrological Processes 15, 2877–2892.

IUSS Working Group, W.R.B., 2006. World reference base for soil resources 2006, 2ndedition. World Soil Resources Reports, 103. FAO, Rome.

Jiménez, C., Tejedor, M., Morillas, G., Neris, J., 2006. Infiltration rate in Andisols: effect ofchanges in vegetation cover (Tenerife, Spain). Journal of Soil and Water Conserva-tion 61, 153–158.

Keith, D.M., Johnson, E.A., Valeo, C., 2010a. A hillslope forest floor (duff) water budgetand the transition to local control. Hydrological Processes 24, 2738–2751.

Keith, D.M., Johnson, E.A., Valeo, C., 2010b. Moisture cycles of the forest floor organiclayer (F and H layers) during drying. Water Resources Research 46.

Klute, A., 1986. Water retention: Laboratory methods. In: Klute, A. (Ed.), Methods ofsoil analysis. Part I. Physical and mineralogical methods. American Society ofAgronomy – Soil Science Society of America, Madison, pp. 635–662.

Lauren, A., Mannerkoski, H., 2001. Hydraulic properties of mor layers in Finland. Scan-dinavian Journal of Forest Research 16, 429–441.

Lauren, A., Mannerkoski, H., Orjasniemi, T., 2000. Thermal and aeration properties ofmor layers in Finland. Scandinavian Journal of Forest Research 15, 433–444.

Letey, J., 1969. Measurement of contact angle, water drop penetration time and criticalsurface tension. Proceedings of the Symposium onWater Repellent Soils, Riverside,EEUU, pp. 43–47.

Lin, C.Y., Chou, W.C., Tsai, J.S., Lin, W.T., 2006. Water repellency of Casuarina wind-breaks (Casuarina equisetifolia forst.) caused by fungi in central Taiwan. EcologicalEngineering 26, 283–292.

Maeda, T., Soma, K., 1985. Classification of Andisols in Japan based on physical proper-ties. In: Schultz, L.G., van Olphen, H., Mumpton, F.A. (Eds.), International clay con-ference. The Clay Minerals Society, Denver, EEUU, pp. 174–178.

Martínez-Zavala, L., Jordán, A., 2008. Effect of rock fragment cover on interrill soil ero-sion from bare soils in western Andalusia, Spain. Soil Use and Management 24,108–117.

57J. Neris et al. / Catena 108 (2013) 50–57

Martínez-Zavala, L., Jordán-López, A., 2009. Influence of different plant species onwater repellency in Mediterranean heathland soils. Catena 76, 215–223.

Martínez-Zavala, L., González, F.A., Jordán, A., 2009. Intensity and persistence of waterrepellency in relation to vegetation types and soil parameters in MediterraneanSW Spain. Geoderma 152, 361–374.

Marzol Jaén, M.V., 2005. La captación del agua de la niebla en la isla de Tenerife. Seviciode Publicaciones de la Caja General de Ahorros de Canarias. 220 pp.

Mataix-Solera, J., Cerdà, A., Arcenegui, V., Jordán, A., Martínez-Zavala, L., 2011. Fireeffects on soil aggregation: a review. Earth-Science Reviews 109, 44–60.

Molina, A., Govers, G., Vanacker, V., Poesen, J., Zeelmaekers, E., Cisneros, F., 2007. Runoffgeneration in a degraded Andean ecosystem: interaction of vegetation cover andland use. Catena 71, 357–370.

Nacci, S., Pla, I., 1991. Técnicas y equipos desarrollados en el país para evaluar propie-dades físicas de los suelos. FONAIAP, Maracay. 48 pp.

Nanzyo, M., Shoji, S., Dahlgren, R., 1993. Physical characteristics of volcanic ash soils. In:Shoji, S., Nanzyo, M., Dahlgren, R. (Eds.), Volcanic ash soils: genesis, properties andutilization. Elsevier Science Publishers B.V., Amsterdam, p. 288.

Perret, S., Michellon, R., Boyer, J., Tassin, J., 1996. Soil rehabilitation and erosion controlthrough agro-ecological practices on Reunion Island (French Overseas Territory,Indian Ocean). Agriculture, Ecosystems & Environment 59, 149–157.

Perrin, J.L., Bouvier, C., Janeau, J.L., Menez, G., Cruz, F., 2001. Rainfall/runoff processes ina small peri-urban catchment in the Andes mountains. The Rumihurcu Quebrada,Quito (Ecuador). Hydrological Processes 15, 843–854.

Podwojewski, P., Poulenard, J., Zambrana, T., Hofstede, R., 2002. Overgrazing effects onvegetation cover and properties of volcanic ash soil in the páramo of Llangahuaand La Esperanza (Tungurahua, Ecuador). Soil Use and Management 18, 45–55.

Poulenard, J., Podwojewski, P., Janeau, J.L., Collinet, J., 2001. Runoff and soil erosionunder rainfall simulation of Andisols from the Ecuadorian Páramo: Effect of tillageand burning. Catena 45, 185–207.

Poulenard, J., Michel, J.C., Bartoli, F., Portal, J.M., Podwojewski, P., 2004.Water repellency ofvolcanic ash soils fromEcuadorian Paramo: effect of water content and characteristicsof hydrophobic organic matter. European Journal of Soil Science 55, 487–496.

Quantin, P., 1994. The Andosols. 15thWorld Congress of Soil Science, Acapulco, Mexico, 6,pp. 848–859.

Regalado, C.M., Ritter, A., 2009. A bimodal four-parameter lognormal linear model ofsoil water repellency persistence. Hydrological Processes 23, 881–892.

Robichaud, P.R., Waldrop, T.A., 1994. A comparison of surface runoff and sediment yieldsfrom low-severity and high-severity site preparation burns.Water Resources Bulletin30, 27–34.

Rodríguez Rodríguez, A., Guerra, J.A., Gorrín, S.P., Arbelo, C.D., Mora, J.L., 2002. Aggre-gates stability and water erosion in Andosols of the Canary Islands. Land Degrada-tion & Development 13, 515–523.

Sharma, K.D., Singh, H.P., Pareek, O.P., 1983. Rainwater infiltration into a bare loamysand. Hydrological Sciences Journal 28, 417–424.

Soil Survey Division Staff, 1993. Soil survey manual. Soil Conservation Service. Hand-book, 18. U.S. Department of Agriculture. 437 pp.

Soil Survey Staff, 1999. Soil Taxonomy. A basic system of soil classification for makingand interpreting soil surveys. N.R.C.S. USDA, Washington, D.C. 869 pp.

Tejedor, M.L., Hernández-Moreno, J.M., Jiménez, C.C., 2007. Soils of volcanic systemsin Spain. In: Arnalds, O., et al. (Ed.), Soils of volcanic regions in Europe. Springer,Berlin, pp. 101–112.

Walkley, A., Black, I.A., 1934. An examination of the Degtjareff method for determiningsoil organic matter, and a proposed modification of the chromic acid titrationmethod. Soil Science 37, 29–38.

Wallach, R., Jortzick, C., 2008. Unstable finger-like flow in water-repellent soils duringwetting and redistribution – the case of a point water source. Journal of Hydrology351, 26–41.

Warkentin, B.P., Maeda, T., 1980. Physical and mechanical characteristics of Andisols.In: Theng, B.K.G. (Ed.), Soils with variable charge. Offset Publications, PalmerstonNorth, pp. 281–352.

Woods, S.W., Balfour, V.N., 2008. The effect of ash on runoff and erosion after a severeforest wildfire, Montana, USA. International Journal of Wildland Fire 17, 535–548.

Woods, S.W., Balfour, V.N., 2010. The effects of soil texture and ash thickness on thepost-fire hydrological response from ash-covered soils. Journal of Hydrology 393,274–286.

Zavala, L.M., Jordán, A., Gil, J., Bellinfante, N., Pain, C., 2009. Intact ash and charred litterreduces susceptibility to rain splash erosion post-wildfire. Earth Surface Processesand Landforms 34, 1522–1532.

Zehetner, F., Miller, W.P., 2006. Erodibility and runoff-infiltration characteristics of vol-canic ash soils along an altitudinal climosequence in the Ecuadorian Andes. Catena65, 201–213.