Effects of moisture conditions on potential soil water repellency in a tropical forest regenerated after ﬁ re

(1)

Effects of moisture conditions on potential soil water repellency in a tropical forest regenerated after ﬁ re

Masako Kajiura

^a,

⁎

^,1

, Takeshi Tokida

^b,1

, Katsutoshi Seki

^b,2

aDepartment of Forest Science, Graduate School of Agricultural and Life Sciences, University of Tokyo, 1-1-1 Yayoi, Bunkyo-ku, Tokyo 113-8657, Japan

bDepartment of Biological and Environmental Engineering, Graduate School of Agricultural and Life Sciences, University of Tokyo, 1-1-1 Yayoi, Bunkyo-ku, Tokyo 113-8657, Japan

a b s t r a c t a r t i c l e i n f o

Article history:

Received 9 April 2011

Received in revised form 19 February 2012 Accepted 27 February 2012

Available online 31 March 2012 Keywords:

Soil water repellency Water content Organic matter

Water-extractable organic matter Tropical forest

Potential water repellency (PWR) is a common index to indicate the degree to which soils repel water. Soil organic matter (SOM) is a requisite substance for the water repellency but many studies have shown that SOM content alone could not fully account for the observed variation in PWR. We investigated potential factors responsible for PWR of soils in a tropical forest in East Kalimantan, Indonesia. In addition to the well- investigated soil properties (e.g., total SOM content, pH, and the amount of iron or aluminum oxides), we also focused on soil moisture content at the time of sampling and water-extractable organic matter (WEOM) content based on the hypothesis that PWR may depend on amphiphilic fractions, including WEOM, in the outermost layer of SOM adsorbed on soil particles—soil water may change the amount and/or the conformation of the amphiphilic fractions. Results showed that the degree of PWR had the highest correlation with the amount of WEOM, not with SOM, among the factors investigated (R²= 0.29). The WEOM content (R²= 0.65) better explained the variation in water repellency than SOM content (R²= 0.47) even after soils were soaked inn-hexane (a non-polar solvent) and expected to have uniform SOM conformation (hydrophobic components dominant on the surfaces). The combination of soil moisture and SOM contents better explained the PWR than SOM content alone. These results suggest that soil water content can have substantial effects on PWR by changing the availability and/or conformation of the amphiphilic SOM, including WEOM.

1. Introduction

Soil water repellency affects water movement and associated nutrient cycles. High soil water repellency reduces the water storage capacity of the soil and enhances the spatial heterogeneity of water and of nutrient movements (Bundt et al., 2001; Doerr et al., 2000;

Kobayashi and Shimizu, 2007). Potential water repellency (PWR), i.e., the water repellency when the soil is dry, has been commonly used to compare the degree to which different soils repel water.

PWR has been reported to increase with increasing soil organic matter (SOM) content if the other soil physicochemical properties are similar (Chenu et al., 2000; Mataix-Solera and Doerr, 2004;

McKissock et al., 1998). Soil water repellency results from the coating of hydrophilic mineral particles with SOM, which is fairly hydrophobic (Doerr et al., 2000). Soil with a high SOM content has a larger

proportion of SOM-covered particle surfaces and is therefore more capable of repelling water.

In addition to the amount of SOM, the conformation of the outermost layer of SOM can signiﬁcantly inﬂuence the PWR. Amphiphilic compounds may be present in the outermost layer of SOM and play a key role in water repellency (Kleber et al., 2007). Amphiphilic compounds can easily change their conformation if soaked in solution, depending on the solution polarity. Soil water repellency has been found to decrease when the soil is dried after soaking in polar solvents, but it increases when soaked in nonpolar solvents (Ma'shum and Farmer, 1985; Ma'shum et al., 1988; Roy and McGill, 2000). Non- polar solvents may change the conformation of the amphiphilic SOM by causing the hydrophobic components to be reoriented outward or stretched on the surface of the soil particles (Doerr et al., 2005;

Ma'shum and Farmer, 1985; Ma'shum et al., 1988). In addition, nonpolar solvents may reorient hydrophilic functional groups inward toward mineral surfaces or toward the inner hydrophilic parts of the adsorbed SOM.

The very high polarity of water may reduce PWR by changing the conformation of the outermost layer of SOM (Ma'shum and Farmer, 1985; Ma'shum et al., 1988). Furthermore, soil moisture conditions may affect PWR by changing the quantity and quality of the SOM, especially water-extractable (soluble) organic matter (WEOM),

⁎ Corresponding author. Tel.: + 81 29 838 8327; fax: + 81 29 838 8199.

E-mail address:kajico@affrc.go.jp(M. Kajiura).

1Present address: National Institute for Agro-Environmental Sciences, 3-1-3 Kannondai, Tsukuba, Ibaraki, 305-8604, Japan.

2Present address: Faculty of Business Administration, Toyo University, 5-28-20 Hakusan, Bunkyo-ku, Tokyo, 112-8606, Japan.

doi:10.1016/j.geoderma.2012.02.028

Contents lists available atSciVerse ScienceDirect

Geoderma

j o u r n a l h o m e p a g e : w w w . e l s e v i e r . c o m / l o c a t e / g e o d e r m a

(2)

through leaching and/or microbiological reactions (Kalbitz et al., 2000). However, little attention has been paid to soil moisture conditions when investigating PWR, although soil water content is a major determinant of actual water repellency (i.e., the changes in soil water repellency as the soil water content varies) (de Jonge et al., 2007;

Kajiura et al., 2012; Kawamoto et al., 2007; Regalado and Ritter, 2005).

Potential water repellency is also affected by the soil pH, soil tex- ture, and Fe-oxide content. Soils with higher pH tend to have lower water repellency (Mataix-Solera et al., 2007). The presence of clay or iron oxides decreases water repellency by increasing the hydrophilic mineral surfaces (Harper and Gilkes, 1994; Harper et al., 2000; McKissock et al., 2002).

In this study, we attempted to identify factors responsible for PWR in tropical forest soils in East Kalimantan, Indonesia. The study site had previously been affected by a forestfire, which might have affected soil water repellency by changing the soil's chemical properties (Certini, 2005; DeBano, 2000). For example,fire may decrease SOM content through combustion and increase pH by providing alkali- rich ashes, both of which can lower the soil water repellency (Knicker, 2007; Mataix-Solera et al., 2007). Due to intrinsic andfire- induced spatial variation, soil samples from the site might have variable physicochemical properties, which is useful for investigating the factors responsible for the PWR. We paid special attention to SOM, WEOM, and water conditions at the time of sampling as potential contributors to PWR. Soil moisture conditions may have a greater impact on PWR in the tropics, where SOM turnover is much more rapid and is usually limited by water rather than by temperature, than in other locations. We also used a soaking experiment with a non-polar solvent to investigate whether the SOM conformation can confound the relationship between SOM and PWR.

2. Materials and methods

2.1. Study site and soil sampling

The study site was located in a natural dipterocarp forest in Bukit Bangkirai in East Kalimantan, Borneo, Indonesia. Forests in East Kali- mantan were extensively damaged byﬁres in 1997–1998. At the research site, plots with different burn severities were determined based on the aboveground biomass investigated in January 2001:

heavily damaged (HD), lightly damaged (LD), and unburned (K) (Abe et al., 2002). The severity of the HD and the LD were equivalent toKeeley's (2009)“high-severity burned”and“low-severity burned” categories, respectively. In the HD plot, almost all arboreal vegetation had been burned by theﬁre, and pioneer species such asMacaranga gigantea(Euph.) grew in hollows. Ferns and herbaceous vegetation grew abundantly on a ridge. In the LD plot, arboreal vegetation had been partially burned, stood dead, or had fallen down. The pioneer species M. gigantea(Euph.) dominated, but climax plants such as Shorea parvifolia (Dipt.) and Madhuca kingiana (Sapo.) were also widely distributed. In the HD and LD plots, the vegetation before theﬁre was similar to that of unaffected natural forest in the K plot;

climax species such as Dipterocarpus confertus (Dipt.) and Shorea laeviswere dominant (Abe et al., 2002). Soil sampling was conducted in the three plots. Each plot covered an area of approximately 10,000 m², and the plots were located 500–1000 m apart. Ultisols, which are widely found in the lowland Dipterocarpaceae forest of East Kalimantan, were found in the K and LD plots. In some parts of the K and HD plots, we found layers of white quartz sand, which are quite common in kerangas forest (Ohta and Effendi, 1992; Seki et al., 2010).Seki et al. (2010)identiﬁed the HD, LD, and K soil textures at 0–30 cm depths as sand, sandy clay or sandy clay loam, and sandy clay loam, respectively, according to the system of the International Union of Soil Science.

To assess variability in soil properties among plots and at different depths, we collected disturbed soil samples from the surface (3–5 cm depth) and subsurface layers (4.5–20-cm depth, with thickness of 3–8 cm for each sample) at 3, 4, and 3 points in the HD, LD, and K plots, respectively, in August 2006. The horizontal sampling area was 15 cm × 15 cm. The surface and subsurface horizons were distin- guished visually by the soil brightness. We packed the samples in plastic bags and stored them in a refrigerator until they were analyzed for SOM and WEOM content, soil water content at the time of sampling, PWR, pH(H2O), electrical conductivity (EC), and iron and aluminum oxide content.

In August 2007, we collected soil samples at 19 different points at 0–25 cm depths and 5-cm intervals using an auger (ϕ4 cm × 50 cm, DK-106B; Daiki Rika Kogyo Co., Ltd., Saitama, Japan). Five of these points were located in the HD plot and 14 in the LD plot, and all sampling points were separated by at least 20 m. We measured the soil water content, SOM content, and PWR of all 0–5- and 5–10-cm samples and some deeper samples (10–25 cm). The samples from both years (2006, 2007) and different plots (HD, LD, K) and depths (surface, subsurface) had varying properties, which helped us determine the underlying mechanisms behind the observed PWR variation (Table 1).

2.2. Soil water repellency

We measured water repellency using the molarity of ethanol droplets (MED) test (King, 1981) and the water drop penetration time (WDPT) test (Letey, 1969). For the initial assessment of PWR, we adopted the MED test for all samples. We placed droplets of ethanol solutions (0–5 M in 0.2 M intervals) onflattened subsample soil surfaces using pipettes. We recorded the MED value for each sample, i.e., the lowest molarity ethanol solution that was able to penetrate the soil surface within 10 s. The smaller the MED value is, the lower the water repellency. For the samples with MED = 0.0 M, we conducted the WDPT test. We placedfive droplets of deionized water onflattened subsample surfaces with a pipette and measured the time for complete infiltration. We recorded the mean value of the times forfive droplets as the WDPT value. For soils with a MED value of 0.0 M, the WDPT test is useful because it subdivides low water repellency using a range of 0–10 s.

To maximize the value extracted from all of our samples, we categorized the water repellency of every sample into one of 4 groups:

strongly water repellent (MED≥1.0 M), moderately water repellent (0.2 M≤MEDb1.0 M), slightly water repellent (1 sbWDPT≤10 s), and wettable (WDPT≤1 s). The four classes were treated as an ordinal categorical variable, not a quantitative variable.

2.3. Soil chemical properties

The soil water content was determined gravimetrically by drying samples for 2 days at 105 °C. Except for the measurements of WEOM, pH(H2O), and EC, an air-dried and sieved (b2 mm) sample was used for analysis. Total carbon content was measured using an NCS analyzer (NA 1500; Carlo Erba Instruments, Milan, Italy) after re- moving visible roots and grinding the air-dried samples tob250μm.

The carbon content reﬂects the SOM content because the soils were not calcareous. The soil pH(H2O) was measured in 1:2.5 soil:water suspensions, and the EC was measured in 1:5 soil:water suspensions using a glass electrode (CM40-S; DKK-TOA Corp., Tokyo, Japan). Free iron oxides (Fed) and amorphous and quasi-crystalline aluminum oxides (Alo) were extracted using a dithionite-citrate (DC) solution and an ammonium oxalate solution, respectively (Mehra and Jackson, 1960; Tamm, 1922). The Fed and Alo contents were quantiﬁed using an atomic absorption spectrophotometer (Z-1600; Hitachi High-Technologies Corp., Tokyo, Japan). WEOM was extracted using 1:10 soil:water suspensions. We added water to undried soils,

(3)

shook them for 30 min, separated the supernatants by centrifugation at 8000 rpm, andfiltered them with 0.45-μm PTFEfilters. We measured the carbon content in thefiltrates, a proxy of WEOM content, using a TOC analyzer (TOC-VCSH/CSN; Shimadzu Corp., Kyoto, Japan).

2.4. Effect of SOM conformation on water repellency

We used a nonpolar organic solvent,n-hexane, to evaluate the effect of the SOM conformation on water repellency.n-Hexane has low polarity (its relative permittivity is 1.9 at 25 °C). It only extracts very hydrophobic SOM (Corradini et al., 2006). It is less toxic than chloro- form, which is often used with water-repellent soils for extraction or for soaking experiments. We put air-dried soils andn-hexane into 50 ml plastic tubes to attain a 1:1 soil–hexane ratio (dry weight), sealed the tubes tightly, and shook them gently a few times. After 48 h at room temperature, we ﬁltered the contents of the tubes using qualitative ﬁlters (No. 2; Tokyo Roshi Kaisha, Ltd., Tokyo, Japan). The soaking time was based onMa'shum et al. (1988). We air-dried the residues for approximately 2 days, stirred them gently, and measured water repellency using the MED test.

2.5. Data analysis

We evaluated the potential correlates of PWR with regression analysis. A general linear model (PROC GLM of SAS v.9.2, SAS Institute Inc.) was used. For categorical PWR (Section 2.2), we applied the multinomial model for the regression analysis (Dobson, 2002). Calcula- tions were performed using the GENMOD procedure of SAS with the cumulative logit link function.

We also evaluated whether theﬁre and soil depth affected soil chemical properties using a mixed-effects model. To this end, the

“ﬁre”effects (differences among HD, LD, and K plots), soil depth or

“horizon”(difference between surface and subsurface), and their interactions (fire × horizon) were treated asfixed effects. The“horizon” was nested within“fire”in the statistical design. The “location” of each plot was designated as a random factor. Sampling points within each plot served as replications (i.e., sampling points were assumed to be independent of each other) to evaluate thefire effects (n = 3, 4, and 3 for HD, LD and K, respectively). Ideally, spatial correlations would be checked before applying ANOVA, but insufficient data were available. Our statistical results should thus be interpreted under the consideration of limited spatial replications. For computa- tion, the PROC MIXED of SAS v9.2 with the restricted maximum likelihood (REML) method was used (Littell et al., 2006). The Kenward– Roger method was used to estimate the denominator degree of

freedom (Kenward and Roger, 1997). Post hoc multiple comparisons were also performed with the Bonferroni correction.

3. Results

3.1. Effects ofﬁre and soil depths on soil properties

Soil physicochemical properties varied with sampling depth and plots (Table 1). All soil physicochemical properties but Alo were different between soil depths (“Horizon”), and“ﬁre”affected SOM and WEOM content. We found signiﬁcant interactive effects between

“horizon”and“ﬁre”on SOM, WEOM, and EC. The SOM and WEOM contents were much higher in the K surface samples than those in HD and LD plots, while the differences in subsurface samples were smaller (Table 1).

3.2. Factors affecting PWR

3.2.1. The relationship between soil properties and PWR in soils sampled in 2006

We conducted regression analyses for the 2006 samples with MED > 0. The WEOM content exhibited the highest correlation with PWR (R²= 0.29) of all analyzed soil properties, including total SOM content, moisture content, pH, EC, and Fed and Alo contents. Howev- er, the WEOM explained only a small fraction of the observed variation in PWR. Total SOM content showed a poor correspondence with PWR (R²= 0.14). However, when combined with soil moisture content (ω), the two factors explained the PWR variation well (R²= 0.75,Pb0.05 for both factors). Soil moisture alone did not explain PWR (R²= 0.03).

To determine whether the conformation of organic matter on soil particles confounded the WEOM– or SOM–PWR relationship, we soaked soils inn-hexane. The water repellency was increased by the treatment in most soils, but some samples still showed MED = 0 after treatment. Among samples with MED > 0, the WEOM had a higher correlation with MED (R²= 0.65) than in the pre-treatment samples. Total SOM still showed a lower correspondence with PWR (R²= 0.47) than WEOM.

3.2.2. Combined effects of SOM and soil moisture content on PWR We investigated the contribution of total SOM and soil moisture contents on PWR for samples taken in 2006 and 2007. Lumping together all samples in both years, we found that total SOM content poorly explained the variation in PWR (measured by the MED scale), as found in 2006 samples: i) samples with a wide range of SOM (0.27%

to 3.95%) showed MED= 0 and ii) samples with MED> 0 had a poor Table 1

Potential water repellency (PWR) and related properties of soil samples collected from surface and subsurface layers (“horizon”) of differentﬁre intensities (“ﬁre”) in 2006.

Plot n PWR SOM WEOM pH(H2O) EC Fed Alo

(M) (g-C/g) (μg-C/g) (dS/m) (g/kg) (g/kg)

Surface

HD 3 0.4 (0.5) ab 0.013 (0.002) bc 69 (33) b 4.3 (0.3) ab 0.046 (0.011) ab 0.4 (0.6) a 0.9 (0.8) a

LD 4 0.2 (0.2) ab 0.017 (0.003) b 104 (31) b 4.1 (0.3) ab 0.042 (0.012) abd 3.5 (1.7) a 2.9 (1.4) a

K 3 1.3 (1.0) a 0.030 (0.005) a 243 (119) a 3.8 (0.0) b 0.075 (0.026) ac 3.9 (3.4) a 2.6 (1.7) a

Subsurface

HD 3 0.0 (0.0) b 0.006 (0.004) c 50 (28) b 4.6 (0.3) a 0.030 (0.009) cd 0.7 (1.2) a 1.4 (2.3) a

LD 4 0.0 (0.0) b 0.012 (0.003) bc 59 (15) b 4.2 (0.2) ab 0.034 (0.011) bcd 4.4 (2.1) a 3.8 (1.4) a

K 3 0.0 (0.0) b 0.010 (0.005) bc 75 (20) b 4.2 (0.4) ab 0.041 (0.025) bd 4.6. (4.1) a 2.7 (2.1) a

ANOVA

Fire ns * * ns ns ns ns

Horizon * *** ** ** *** ** ns

Fire × horizon ns ** * ns ** ns ns

SOM, soil organic matter; WEOM, water-extractable organic matter; Fed, free-iron oxides; Alo, amorphous and quasi-crystalline aluminum oxides.

***,Pb0.001; **,Pb0.01; *,Pb0.05; ns, not signiﬁcant. Values in parentheses indicate standard deviations. Different letters indicate signiﬁcant differences (Pb0.05).

(4)

correlation with SOM (R²= 0.15,P= 0.06) (Fig. 1-a). The correlation between SOM and water repellency remained poor after the hexane treatment (R²= 0.098 for samples having MED > 0) (Fig. 1-b). Notably, some samples showed MED= 0 after soaking despite containing a large amount of SOM (SOM> 0.02 g g⁻¹).

As found in the 2006 samples, if taken together, two factors (total SOM and soil moisture content,ω) signiﬁcantly contributed to PWR.

Among the samples with MED > 0, the multiple regression yielded the following equation: MED = 0.57 × SOM (%)−0.048 ×ω (%) (R²= 0.30,Pb0.01 for SOM andPb0.05 forω). The equation shows that total SOM and moisture content played opposing roles in PWR;

higher SOM and lower moisture content increased PWR.

To maximize the value extracted from PWR data measured by both MED and WDPT, we assessed the effects of the two factors on PWR by categorizing PWR into four“ordinal”classes (Section 2.2).

At the same water content, increasing SOM led to greater water repellency (Fig. 2). In contrast, the higher the water content was, the lower the PWR expected at a particular SOM content (Fig. 2). Accordingly, there was even an apparent correlation between SOM and ω (r = 0.55, pb0.001); they have opposite effects on PWR. A multinomial logit regression conﬁrmed our conjectures; both SOM andωwere

highly signiﬁcant and contributed to PWR in opposite manners (Table 2). The reproduced categorical PWR as a function of SOM and ωreasonably simulated the observed pattern of PWR (Fig. 2). It is note- worthy that using SOM and ω simultaneously yielded a better ﬁt (AIC = 159) than a simple regression using either SOM (AIC = 211) or ω(AIC= 212) alone.

4. Discussion

4.1. Variability in soil properties

The significant interaction effects between fire and horizon on SOM, WEOM, and EC may reflect the fact that organic matters in surface soils in HD and LD plots were depleted by thefire (Table 1).

Samples from different plots and depths in 2006 and 2007 should be useful for investigating potential factors affecting PWR because we found varying soil properties depending on depth and degree of ﬁre damages.

4.2. Does the conformation of SOM affect the SOM–PWR relationship?

We found a poor correlation between total SOM content and PWR (Fig. 1). One possible explanation is that the conformation of the adsorbed SOM may not have been similar among samples. For example, in samples with similar amounts of SOM, some might have had surfaces rich in hydrophilic functional groups, and in others, the hydrophobic components might have been dominant. If such conformational variability introduces variation into the SOM–PWR relationship, then the hexane treatment would have yielded a better signal by making the conformation more similar among samples (e.g., the surface would have become dominated by the hydrophobic components of SOM). However, the relationship between SOM contents and MED remained poor after soaking in hexane. These results reveal that the conformation of SOM is not a confounding factor and that total SOM content alone is not a good indicator of PWR in the samples investigated in this study.

Fig. 1.Soil water repellency (a) before and (b) after soaking the soils inn-hexane as a function of total soil organic matter (SOM) content. Data from soil samples taken in 2006 and 2007 are shown. Soil water repellency was evaluated with the MED test.

Fig. 2.Observed (open symbols) and model-predicted (colored background) potential water repellency (PWR) plotted against soil water content at the time of sampling (ω) and total soil organic matter content (SOM). Data from 2006 to 2007 are shown. The degree of PWR was evaluated by the MED and WDPT tests and categorized into 4 classes: strongly water repellent (WR) (1.0≤MED (M)), moderately WR (0.2≤MED (M)b 1.0), slightly WR (1bWDPT (s)≤10), and wettable (WDPT (s)≤1). The multinomial logit model estimated the probabilities of the categorized PWR in the SOM-ωrelation- ship (using the‘estimate’values inTable 2). The RGB (red-green-blue) color system shows the probability of each category: strongly, moderately, and slightly WR. Deeper colors show higher probability. The residual black-colored part indicates the high probability of wettable soils.

Table 2

Maximum likelihood parameter estimates from multinomial logit regression: potential water repellency (PWR) vs. soil organic matter content (SOM) and water content (ω).

Parameter^a Estimate Wald statistics (chi-square) Level of signiﬁcance

Intercept 1 −5.10 32.5 ***

Intercept 2 −2.36 17.1 ***

Intercept 3 −0.34 0.45 ns

SOM 3.81 37.7 ***

ω −0.33 38.5 ***

aThe intercept terms corresponded to the three cumulative logits deﬁned on the PWR categories in the following order: i) strongly water repellent (WR), ii) moderately WR, iii) slightly WR, and iv) wettable.

***, Pb0.001; ns, not signiﬁcant.

(5)

4.3. WEOM as a factor for PWR

The poor correspondence between PWR and SOM might be caused by the dependence of PWR on the outermost layer of the SOM, not on the total SOM adsorbed on soil particles.Doerr et al. (2005)found that the amount of the amphiphilic SOM (up to 25% of total SOM) extracted by an isopropanol/ammonium mixture showed a very weak relationship with PWR, although the solvent fully diminished the water repellency. We thus propose that only a fraction of the amphiphilic compounds might be critical for PWR. The outermost layer of the SOM may consist of unstable fractions that are loosely retained on the surface of soil particles (Kleber et al., 2007) and could be extracted by water.

This conjecture agrees well with theﬁnding obtained from the 2006 samples; WEOM is the best indicator for PWR among all parameters investigated. The high WEOM-MED correlation after then-hexane soaking also shows that the WEOM could explain a major component of MED variation when the effects of WEOM conformation were removed (Section 3.2.1). Thus, it is likely that WEOM represents the crucial fraction of SOM with respect to PWR at this study site. Our results are consistent with reports ofFranco et al. (2000) showing that extracted WEOM can cause water repellency.

4.4. Enigmatic but signiﬁcant effect of soil water content on PWR The soil moisture content at the time of sampling signiﬁcantly re- duced the PWR, as shown by the linear and multinomial logit regression analyses (Section 3.2.2,Fig. 2). Soil water might wash away the WEOM, change its conformation (make hydrophilic parts dominant in soil particle surfaces), and/or promote the decomposition of WEOM (Doerr et al., 2005; Kalbitz et al., 2000; Ma'shum and Farmer, 1985; Ma'shum et al., 1988). If WEOM is the major determinant of PWR, as suggested above, these processes could decrease PWR. We are aware of no previous studies investigating the effects ofωonpotentialwater repellency, whereasactualwater repellency has been shown to depend on soil water conditions (de Jonge et al., 2007; Kajiura et al., 2012; Kawamoto et al., 2007; Regalado and Ritter, 2005). Clearly and most importantly, our results call for addi- tional studies on soil moisture content as a potential factor regulating WEOM, which may in turn affect PWR.

5. Conclusions

We investigated the factors responsible for PWR in tropical forest soils in East Kalimantan, Indonesia. We found that the total SOM content only partially explained the variation in PWR and that the combination of the SOM content andω, the moisture content at the time of soil sampling, explained this phenomenon better. Theﬁndings of this study suggest that PWR might change frequently due to rainfall. The WEOM content is sensitive to soil moisture and might therefore directly control PWR. However, future investigations are needed to conﬁrm this conjecture.

Acknowledgments

This study was supported by the Biodiversity and Ecosystem Restoration Program, 21st Century Center of Excellence Program at the University of Tokyo, and the Global Environment Research Fund (E-051) of the Ministry of the Environment, Japan. We thank Kaori Suzuki, Satsuki Kodama, Chihiro Kato, Noriyuki Kamachi, Yoshiki Shimowada, Taku Nishimura, Tsuyoshi Miyazaki, the staff at Bukit Bangkirai, and the late Dr. Herwint Simbolon for support during the ﬁeld investigations.

References

Abe, Y., Hirata, Y., Shimizu, H., 2002. E-2 Studies on the impacts of forestﬁres on natural resources and on the evaluation of ecosystems. Final Reports on Environment Research & Technology Development Fund, pp. 1–278 (in Japanese with English abstract).

Bundt, M., Jaggi, M., Blaser, P., Siegwolf, R., Hagedorn, F., 2001. Carbon and nitrogen dynamics in preferentialﬂow paths and matrix of a forest soil. Soil Science Society of America Journal 65 (5), 1529–1538.

Certini, G., 2005. Effects ofﬁre on properties of forest soils: a review. Oecologia 143 (1), 1–10.

Chenu, C., Le Bissonnais, Y., Arrouays, D., 2000. Organic matter inﬂuence on clay wetta- bility and soil aggregate stability. Soil Science Society of America Journal 64 (4), 1479–1486.

Corradini, D., Grego, S., Liburdi, K., Marinari, S., Tonci, M., 2006. A combined approach employing soxhlet extraction and linear gradient elution reversed-phase HPLC for theﬁngerprinting of soil organic matter according to hydrophobicity. Chroma- tographia 63, S11–S16.

de Jonge, L.W., Moldrup, P., Jacobsen, O.H., 2007. Soil-water content dependency of water repellency in soils: Effect of crop type, soil management, and physical- chemical parameters. Soil Science 172 (8), 577–588.

DeBano, L.F., 2000. Water repellency in soils: a historical overview. Journal of Hydrology 231–232, 4–32.

Dobson, A.J., 2002. An introduction to generalized linear models, Second Edition. Chapman and Hall/CRC Press, Washington, D.C., USA.

Doerr, S.H., Shakesby, R.A., Walsh, R.P.D., 2000. Soil water repellency: its causes, char- acteristics and hydro-geomorphological signiﬁcance. Earth-Science Reviews 51 (1–4), 33–65.

Doerr, S.H., Llewellyn, C.T., Douglas, P., Morley, C.P., Mainwaring, K.A., Haskins, C., Johnsey, L., Ritsema, C.J., Stagnitti, F., Allinson, G., Ferreira, A.J.D., Keizer, J.J., Ziogas, A.K., Diamantisf, J., 2005. Extraction of compounds associated with water repellency in sandy soils of different origin. Australian Journal of Soil Research 43 (3), 225–237.

Franco, C.M.M., Clarke, P.J., Tate, M.E., Oades, J.M., 2000. Hydrophobic properties and chemical characterisation of natural water repellent materials in Australian sands. Journal of Hydrology 231–232, 47–58.

Harper, R.J., Gilkes, R.J., 1994. Soil attributes related to water repellency and the utility of soil survey for predicting its occurrence. Australian Journal of Soil Research 32 (5), 1109–1124.

Harper, R.J., McKissock, I., Gilkes, R.J., Carter, D.J., Blackwell, P.S., 2000. A multivariate framework for interpreting the effects of soil properties, soil management and landuse on water repellency. Journal of Hydrology 231–232, 371–383.

Kajiura, M., Etori, Y., Tange, T., 2012. Water condition control of in situ soil water repellency: an observational study from a hillslope in a Japanese humid-temperate forest. Hydrol. Process. doi:10.1002/hyp. 8310.

Kalbitz, K., Solinger, S., Park, J.H., Michalzik, B., Matzner, E., 2000. Controls on the dynamics of dissolved organic matter in soils: a review. Soil Science 165 (4), 277–304.

Kawamoto, K., Moldrup, P., Komatsu, T., de Jonge, L.W., Oda, M., 2007. Water repellency of aggregate size fractions of a volcanic ash soil. Soil Science Society of America Journal 71 (6), 1658–1666.

Keeley, J.E., 2009. Fire intensity,ﬁre severity and burn severity: a brief review and suggested usage. International Journal of Wildland Fire 18 (1), 116–126.

Kenward, M.G., Roger, J.H., 1997. Small sample inference forﬁxed effects from restricted maximum likelihood. Biometrics 53 (3), 983–997.

King, P.M., 1981. Comparison of methods for measuring severity of water repellence of sandy soils and assessment of some factors that affect its measurement. Australian Journal of Soil Research 19 (4), 275–285.

Kleber, M., Sollins, P., Sutton, R., 2007. A conceptual model of organo-mineral interactions in soils: self-assembly of organic molecular fragments into zonal structures on mineral surfaces. Biogeochemistry 85 (1), 9–24.

Knicker, H., 2007. How doesﬁre affect the nature and stability of soil organic nitrogen and carbon? A review. Biogeochemistry 85 (1), 91–118.

Kobayashi, M., Shimizu, T., 2007. Soil water repellency in a Japanese cypress plantation restricts increases in soil water storage during rainfall events. Hydrological Processes 21 (17), 2356–2364.

Letey, J., 1969. Measurement of contact angle, water drop penetration time, and critical surface tension. Symposium on water-repellent soils, University of California, Riverside, CA, pp. 43–47.

Littell, R.C., Milliken, G.A., Stroup, W.W., Wolﬁnger, R.D., Schabenberger, O., 2006. SAS®

for Mixed Models, 2nd edition. SAS Institute Inc., Cary, NC.

Ma'shum, M., Farmer, V.C., 1985. Origin and assessment of water repellency of a sandy south Australian soil. Australian Journal of Soil Research 23 (4), 623–626.

Ma'shum, M., Tate, M.E., Jones, G.P., Oades, J.M., 1988. Extraction and characterization of water-repellent materials from Australian soils. Journal of Soil Science 39 (1), 99–110.

Mataix-Solera, J., Doerr, S.H., 2004. Hydrophobicity and aggregate stability in calcareous topsoils fromﬁre-affected pine forests in southeastern Spain. Geoderma 118 (1–2), 77–88.

Mataix-Solera, J., Arcenegui, V., Guerrero, C., Mayoral, A.M., Morales, J., Gonzalez, J., Garcia-Orenes, F., Gomez, I., 2007. Water repellency under different plant species in a calcareous forest soil in a semiarid Mediterranean environment. Hydrological Processes 21 (17), 2300–2309.

McKissock, I., Gilkes, R.J., Harper, R.J., Carter, D.J., 1998. Relationships of water repellency to soil properties for different spatial scales of study. Australian Journal of Soil Research 36 (3), 495–507.

(6)

McKissock, I., Gilkes, R.J., Walker, E.L., 2002. The reduction of water repellency by added clay is inﬂuenced by clay and soil properties. Applied Clay Science 20 (4–5), 225–241.

Mehra, O., Jackson, M., 1960. Iron oxide removal from soils and clays by a dithionite-citrate system buffered with sodium bicarbonate. Clays and Clay Minerals 7 (1), 317–327.

Ohta, S., Effendi, S., 1992. Ultisols of lowland dipterocarp forest in East Kalimantan, In- donesia. Soil Science and Plant Nutrition 38, 197–206.

Regalado, C.M., Ritter, A., 2005. Characterizing water dependent soil repellency with minimal parameter requirement. Soil Science Society of America Journal 69 (6), 1955–1966.

Roy, J.L., McGill, W.B., 2000. Flexible conformation in organic matter coatings: an hypothesis about soil water repellency. Canadian Journal of Soil Science 80 (1), 143–152.

Seki, K., Suzuki, K., Nishimura, T., Mizoguchi, M., Imoto, H., Miyazaki, T., 2010. Physical and chemical properties of soils in theﬁre-affected forest of East Kalimantan, Indonesia.

Journal of Tropical Forest Science 22 (4), 414–424.

Tamm, O., 1922. Eine methode zur bestimmung der anorganischen komponenten des gelkomplexes im boden. Meddelelser Statens Skogsförsöksanst 19, 387–404.