Effects of vegetation cover on sediment particle size distribution and transport processes in natural rainfall conditions on post-fire hillslope plots in South Korea.
In South Korea, forest fires mostly occur in March and April, and the mean annual frequency and intensity of occurrence of forest fires has increased because of human-induced fires as a consequence of more people increasingly enjoying forest-based outdoor recreation activities. From 1970 to 2003, forest fires occurred, on average, 429 times per year and the mean burned area was 2908 ha [year.sup.-1] (Lee et ai 2006), compared with 460 times per year and a mean burned area of 832 ha [year.sup.-1] from 2005 to 2009 (Lim et al. 2010), causing aggravated soil erosion, landslides, sediment runoff and ecological damage in the burned areas. Several authors have studied post-fire responses of natural vegetation recovery in South Korea (Choung et al. 2004; Kim et al. 2008) and the erosion, runoff and surface soil stability response to vegetation recovery from various post-wildfire rehabilitation treatments at various spatial and temporal scales (Shin et al. 2013; Lee et ai 2014).
Rainfall erosion involves the detachment and transport of surface soil particles by raindrop impact and runoff energy associated with rainsplash and rainflow processes in interrill erosion (Gabet and Dunne 2003), as well as dry ravel mechanisms, such as wind, gravity or animal activity, on burned hillslopes (Shakesby and Doerr 2006; Moody 2010; Moody et al. 2013). Rain splash transport depends on the kinetic energy or momentum of raindrop impact detachment related to rainfall intensity, and is particularly effective and most important on burned hillslopcs with patchy areas of bare, exposed soil void of litter, duff and vegetation cover (Moody 2010; Moody et al. 2013). Thus, post-fire hillslopcs experience marked erosion and sediment runoff transport processes following increased bare ground exposed and low soil structural stability that allows aggregate breakdown by increased raindrop impact and runoff force immediately after the fire (Cerda 1998; Benavides-Solorio and MacDonald 2005; Larsen et al. 2009; Moody et al. 2013).
However, the amount of sediment movement is reduced 4 years after a forest fire due to partial vegetation recovery in the landscape and declining sediment availability after numerous post-fire storm events (Prosser and Williams 1998). Therefore, on burned hillslopes undergoing appreciable vegetation and litter cover recovery from seeding and/or sprouting herbs and shrub growth, the soil cover providing surface roughness and aggregate stability becomes a dominant controlling factor of post-fire soil erosion and aggregate size distribution. In addition to vegetal cover, rainfall-runoff energy, slope, fire severity, soil structure and aggregate stability arc dominant controlling factors of soil erosion and runoff in post-fire hillslopes (Cerda and Doerr 2005; Shakesby and Doerr 2006; Mataix-Solera et al. 2011; Shakesby 2011). On undisturbed forested hillslopes, soil erosion is controlled by both canopy and ground vegetation cover, plant root soil and organic matter binding and aggregate stabilisation activities on the forest floor that significantly reduce raindrop impact and runoff energy, as well as sediment detachment and transport.
Sediment particle size distribution characteristics related to erosion and transport mechanisms and vegetal cover have been mostly studied using simulated rainfall experiments in field or laboratory cultivated soils (Young 1980; Meyer et al. 1992; Farenhorst and Bryan 1995). However, natural rainfall experiments in the field also provide a platform to investigate the particle size distribution of eroded sediment in relation to sediment detachment and transport mechanisms (Slattery and Burt 1995; Martinez-Mena et al. 1999; Parsons et al. 2006). The size distribution of eroded sediment is reported to be affected by several factors, including the particle size distribution of the original soil, aggregate breakdown during erosion and the settling velocity of different size classes of particles or aggregates (Loch and Donnollan 1983; Proffitt and Rose 1991). When adequate structural aggregate breakdown occurs due to raindrop impact and runoff energy on bare soils, the particle size of eroded sediments in runoff can be expected to be coarser than that of the uncroded soil. If no breakdown in the soil structure occurs during rainfall erosion, the assumption is that the particle size distribution at the steady state will be the same as that of the uncroded soil because eroded sediment has a degree of aggregation equal to or less than that of the in situ soil (Hairsine et al. 1999).
'Aggregate stability' refers to soil resistance in maintaining its own structure when it is subjected to external disruptive forces of dispersion (e.g. from the impact of raindrop and water erosion existing in the field) and, together with size distribution, can be used as an indicator of the state of soil structure, physical stability and erodibility related to field phenomena (Kemper and Roscnau 1986; Mataix-Solera et al. 2011). In particular, high-severity fires cause disaggregation and reduce aggregate stability in sandy soils that have organic matter as the principal binding agent as a consequence of organic matter destruction, enforcing soil crusting and surface sealing, which reduces water infiltration and increases soil erodibility (Mataix-Solera et al. 2011).
The strength of aggregate stability and the size distribution of the soil, the characteristics of the erosion event affecting the extent of soil aggregate breakdown and the characteristics of the eroded sediment particle size and their mean weight diameter (MWD) may be quantified in terms of the amount of rainfall and runoff energy and their effects on aggregate breakdown (Teixeira and Misra 1997). In addition, the size distribution of transported sediments has been studied using the enrichment ratio (ER), which relates the 'effective' and 'ultimate' size distribution of sediments (Ongley et al. 1981), because of increasing evidence that sediments are mostly transported in the form of aggregates rather than primary particles (Miller and Baharuddin 1987; Slattery and Burt 1997; Martinez-Mena et al. 1999). The breakdown of aggregates by raindrop impact and runoff energy is related to percentage vegetal cover, and particle size provides useful information on sediment transport selectivity (Martinez-Mena et al. 1999).
The present study considered the significance of the size distribution of eroded sediment in natural rainfall erosion events in a post-fire hillslope environment of slow vegetation recovery treatment plots (BUP), fast vegetation recovery treatment plots (BSP) and unburned treatment plots (CP). The primary aim of the study was to investigate the effects of vegetation cover, degree of soil aggregate stability, slope and rainfall intensity on the sediment particle size distribution and transport processes on post-fire hillslopes at small plots 6 years after a wildfire.
Materials and methods
Study area and site characteristics
The study area is located in Changpyeong-ri, Chilgok-gun, and extends between latitude 35[degrees]98'12"N and longitude 123[degrees]49'73"E in the Gyeongsangbuk-do province of South Korea (Fig. 1). The upland forested watershed was devastated by a high burn severity wildfire in April 2009 that consumed all trees and ground cover along its path. The area has a humid temperate climate with mean annual rainfall of 1064.4 mm concentrated within the summer months of June, July and August, and average monthly temperature ranges from 23 to 27[degrees]C during these wetter and warmer months, warmer than the national monthly mean (10-16[degrees]C). Total rainfall and intensity arc lower in the study area, with longer dry periods compared with other parts of the country. Vegetation cover in the post-fire hillslope plots was spatially variable and consisted of planted grass and naturally recovered native herbs and shrubs. Dominant plant species that regenerated at the burned hillslope plots planted with P. densiflora were Quercus spp., Lespedeza cyrtobotrya,
Smilax China. Carex lanceolata, and Pteridium aquilinum var. latiusculum. Dominant understory plant species at the unbumed plots with mature P. densiflora were by Quercus spp.. Zanthoxylum spp.. Lespedeza cyrtobotiya, Smilax China, Rhus trichocarpa, Ligustrum obtusifolium. Pteridium aquilinum var. latiusculum. Robinia pseudoacacia, and Rhododendron spp. Vegetation cover did not reach an expected minimum of 60% cover in the slow vegetation recovery hillslope plots 6 years after the wildfire. The different plot-level vegetation conditions were a representation of the post-fire vegetation recovery condition on the study hillslopes.
Erosion plots and experimental design
A total of 15 plots (5 x 2 m, length x width) were constructed with sediment trap structures attached to the base of each plot. The experimental treatment had 10 plots: four replicate plots were located at a slow vegetation recovery ridge hillslope (BUP) and six replicate plots were located at three fast vegetation recovery hillslopcs (BSP). The four plots at the slow vegetation recovery hillslope had comparable total vegetation cover and 70-80% bare soil exposed, with mean slope gradients of 11% for the upper two plots and 22% for the lower two plots. The six plots located at the three fast vegetation recovery hillslopes consisted of two plots at each site with comparable vegetation cover and 20-30% bare soil exposed, with mean slope gradients of 16%, 32% and 46%. For each treatment type, the replicate pairs of plots were located on similar slope gradients, but between-treatment plots had significant differences in vegetation cover to account for the effect of percentage bare soil exposed. The unbumed treatment (CP) had five replicate plots with mean slope gradient of 11 %, 100% combined ground cover of shrubs, vines, herbs, litter and duff layer (1-4 cm depth) and >75% canopy cover, with hardly any visible bare soil exposed in each plot. The plots were located on elevations between 100 and 200 m. Data from a recording tipping bucket rain gauge installed at an open ridge hillslope top were analysed to determine maximum 30-min rainfall intensity (I30), start and end times and the volume and duration of each rainfall event. Vegetation cover was measured in each plot using the point quadrant method in June 2015 during the start of data collection. Soil bum severity was higher in the BUP plots (up to 5 cm soil depth) located at a ridge slope than in the BSP plots (<3 cm soil depth), as reconstructed from the presence of charred roots (0.25-8 cm diameter), organic matter content and the soil structure of the bare soil surface in each plot (Parsons et al. 2010).
Sediment data collection and particle size analysis
Eroded sediments were collected on a rainfall event basis from June to October 2015 and oven dried at 105[degrees]C until all moisture was removed. The dried sediments were sieved with standard test sieves using a multiple flat sieve shaker machine for 3 min, weighed and analysed as the percentage of seven grain size classes of defined limits (<63, 63-125, 125-250, 250-500, 500-1000, 1000-2000 and [greater than or equal to] 2000 [micro]m) present in the
sediments. Sediment samples of uncroded surface soil were also collected from the top 2 cm of soil with a hand spade from the immediate outer boundary after each rainfall event. The collected uneroded sediments were air dried and sorted into the same size classes as the eroded soil sediments without dispersion. The size distribution of the uncroded sediments collected after each rainfall event was compared with that of the eroded sediments of the next event. Soil samples were also collected separately using a 5 x 5 cm standard soil sample core at soil depths of 0-5 and 5-10 cm; these were oven dried to calculate surface soil bulk density, water content and organic matter by the loss on ignition (LOI) method (Heiri et al. 2001). Because the eroded sediment from each CP plot for individual rainfall events was too small to allow for particle size analysis for each plot, the sediments from the five replicate plots were combined in each erosion event to allow for grain size analysis.
Kinetic energy associated with rainfall
The raindrop impact on the bare ground surface is assumed to have a constant terminal velocity of 9.8m [s.sup.-1] and the energy released at the soil surface by raindrop impact and runoff flow per unit area per unit depth is usually estimated as a function of rainfall intensity. Where ground vegetation such as grass, herbs and shrubs provides surface roughness, the impact of raindrop and runoff energy is reduced in interrill erosion. Thus, the effective kinetic energy (ICE) of rainfall for areas covered by ground-level vegetation is calculated using the following equation suggested by Gabet and Dunne (2003):
KE = [pitv.sup.2] (1 - [C.sub.v]) cos [theta]/2
where p is the density of water (assumed to have a constant value of 1000 kg [m.sup.-3] at 25[degrees]C), i is rainfall intensity (m [s.sup.-1]), t is storm duration (s), v is raindrop velocity (assumed to be a constant of 9.8 m [s.sup.-1]), [theta] is hillslope angle and [C.sub.v] is the proportion of the surface area covered by ground-level vegetation and/or canopy vegetation, which intercepts part of the rainfall and reduces the total rainfall force reaching the ground surface as direct throughfall.
Energy associated with runoff
Stream power ([OMEGA]; [W.sup.-1] [m.sup.-2]) is the energy of runoff per unit surface area, some of which may be available to remove and transport aggregates from the erosion surface downslope. Stream power for sheet flow of runoff water is calculated using the equation given by Bagnold (1966) and used in several studies (c.g. Teixeira and Misra 1997; Moody 2010):
[OMEGA] = PGQS
where P is the density of runoff water (assumed to have a constant value of 1000 kg [m.sup.-3]), G is the acceleration due to gravity (assumed to be a constant of 9.8 m [s.sup.-2]), S the sine of the erosion surface slope, and Q is the whole volumetric flux of runoff per unit width of erosion surface ([m.sup.3]/[m.sup.2]/[s.sup.1]). Note that Q is calculated as the volume of runoff (V; [m.sup.-3] [s.sup.-1]) divided by the product of rainfall duration (D) and the width (w) of the plot i.e. Q= V/(D x w).
Data for the eroded and uneroded sediments were analysed by one-way analysis of variance (ANOVA) with Tukey's multiple comparison post hoc tests in SPSS version 21 (IBM). Multiple linear regression analysis was used to determine the relationships between percentage bare soil exposed, slope, bulk density, organic matter, [I.sub.30] and runoff flow as independent variables, with eroded sediment particle size distribution as the dependent variable. Scatter plots, with and without linear regression lines, are used to present the relationships between ER and [I.sub.30] in the BUP, BSP and CP treatments plots.
In the present study, the relative proportion of the eroded and uncroded sediment particle size distribution was used to indicate how the different sediment size classes were eroded and transported by rainfall runoff energy. It is assumed that the eroded sediment has a degree of aggregation equal to or less than that of the uneroded soil (Hairsine et al. 1999). Thus, a comparison of the percentage of particle sizes of the eroded and uneroded sediment (ER) provides a measure of the particle size enrichment and transport selectivity involved in sediment detachment and transport during different natural rainfall intensities, calculated using the following equation (Slattery and Burt 1997; Martinez-Mena et al. 1999):
ER = % Particles in a given size class in surface run-off/ % Particles in a given size class in matrix soil
where ER values >1.0 represent enrichment (i.e. a given class forms a greater proportion of the transported sediment in runoff than in the parent soil) and ER values <1.0 represent depiction (i.e. a given class forms a greater proportion in the matrix soil than in the transported load).
Mean weight diameter
One of the most common methods of analysis and expression of results on soil aggregate size distribution by dry sieving is the MWD (Kemper and Roscnau 1986). The particle size distribution of the eroded sediment was also analysed in the present study using the MWD, calculated from the following summation equation (Kemper and Roscnau 1986; Lc Bissonnais 1996; Teixeira and Misra 1997):
MWD = [7.summation over (i=1)] ([x.sub.i][w.sub.i])
where [x.sub.i] is the mean diameter (pm) of size fraction i, [w.sub.i] is the proportion of the total sample (by dry weight) of the size fraction i and n is the total number of size fractions (i.e. 7 in the present study). The key assumption is that a higher MWD of sediments is related to soil aggregation and lower erosion (Teixeira and Misra 1997).
Results and discussion
Effects of fire on physical properties of soil
In the present study, post-fire soil disaggregation or, conversely, aggregate stability was inferred from measured soil properties such as soil organic matter, soil texture, bulk density and micro aggregate size distribution, as suggested by Mataix-Solera et al. (2011). The soil type in the BUP and BSP treatment plots was sandy soils with 90% and 87.5% sand, 9.8% and 12.3% silt and 0.2% and 0.3% clay content respectively in the top 5 cm of soil depth. However, there were significant differences in the mean gravel content, bulk density, organic matter, porosity and water content (by weight and volume) between the two treatments in the top 5 cm of soil depth (Table 1). The soil type at the CP plots was loamy sand (78.1% sand, 21.5% silt and 0.4% clay).
Significantly lower mean values of gravel content and bulk density, and higher mean values of organic matter, porosity and water content (by weight and volume) were found in the CP compared with the severely burned BUP plots in the top 5 cm of soil depth (Table 1). The results indicate that the high fire severity significantly altered topsoil organic matter in the burned plots. No significant differences were found in soil organic matter, bulk density and water content between the burned and unbumed plots at soil depths of 5-10 cm (data not shown). High fire bum severity can consume organic matter in the top 5 cm of soil depth, reduce structural aggregate stability to loose and single-grain soil particles and cause water rcpellency in the top 10 cm of soil depth (Parsons et al. 2010), whereas moderate burning alters soil properties and related water rcpellency only at soil depths up to 2 cm (Badia-Villas et al 2014).
Effects of vegetation recovery on sediment particle size distribution and aggregate stability
Particle size distribution of the eroded and uncroded sediment showed some significant differences within and between treatments plots (Table 2). The most significant differences within the uncroded and eroded sediments between the treatment plots was found for the finest (<63 [micro]m) and coarsest (> 2000 [micro]m) particle sizes, with CVs of 56.2% and 75.4%, and 48.2% and 102.5% respectively (Table 2). The lowest significant differences within the uncroded and eroded sediments between treatments were found for sand-sized particles between 63 and 2000 [micro]m (Table 2). The results indicate a dissimilar pattern and uneven proportion of eroded sediment particle size distribution between the BUP, BSP and CP treatment plots, mostly attributable to differences in total vegetation cover and soil aggregate stability following rainfall erosion.
The most eroded sediment particle size class for all treatments was that of sand (125-250 [micro]m) and no significant differences were observed between the three treatments plots (P>0.05), mainly due to its availability and susceptibility to transport processes (Moody et al. 2013). The sand particle size class of 125-250 [micro]m represented the highest percentage of sand particles of 63-2000 [micro]m in both the eroded and uncroded sediments in all treatment plots. Young (1980) suggested that particles between 20 and 200 [micro]m are the most crodible because larger particle sizes (>200 [micro]m) have sufficient mass to limit their movement, and cohesive forces prevent the detachment of smaller (< 20 [micro]m) particle sizes, hindering their mobilisation and transport during erosion.
The ER of the median eroded and uncroded particle size distribution showed variable inter-storm contrast between the treatments plots (Fig. 2). The ER of the median diameter size in the BUP plots with average slopes of 11% and 22% was always much greater than 1.0, whereas the ER of median diameter size in the BSP plots (with average slopes of 16%, 32% and 46%) was much closer to 1.0 or less than 1.0 and that in the CP plots (with an average 11% slope) was always less than 1.0 for all individually recorded erosive events from June to October (Fig. 2).
The coarser median diameter size of the eroded than uneroded sediment in the BUP plots indicates that most of the eroded sediment included a substantial proportion of macroaggregates and gravel-sized particles because of considerable aggregate breakdown and transport by raindrop and rainflow energy and by wind and gravity dry ravel transport mechanisms. For the BSP and CP treatment plots, the median ER results suggest that most of sediment in the runoff included a substantial proportion micro aggregate (75% and 100%, respectively) due to limited aggregate breakdown and sediment transport dominated by rainsplash. The former result is contrary to, whereas the latter is in agreement with the findings of previous studies (Slattery and Burt 1997; Martinez-Mena et al. 1999).
The ER of sand-sized particles (63-2000 [micro]m) was always greater than 1.0, independent of the percentage of bare soil exposed in each treatment plots (Fig. 3a). In the present study, the enrichment of sand particle sizes in all treatment plots reflected the sandy soil texture of the study plots and highlighted the overwhelming influence of soil texture and sand sediment availability on the eroded sediment size distribution in the study area independent of treatment type and rainfall intensity. The ER values of clay silt particle sizes <63 [micro]m were appreciably higher in the BUP plots during some events over time than in the BSP and CP treatment plots (Fig. 36). These results are similar to those of Martinez-Mena et al. (1999), who found ER values of sand-sized particles to be always greater than 1.0 in both disturbed and natural undisturbed plots, but ER values of clay and silt to be mostly <1.0, except in the disturbed plots because of aggregate breakdown.
The enrichment of finer clay silt particles in the BUP plots suggests increased aggregate breakdown by raindrop impact and runoff energy during individual erosive events, especially those of higher intensity. Cohesive forces prevented the detachment of smaller particles <63 [micro]m and hindered their mobilisation and transport during erosion in the BSP and CP plots than in the BUP plots. This result conforms to the common observations that soil aggregate stability is positively associated with undisturbed vegetation, litter and organic matter cover (Albaladejo et al. 1998; Martinez-Mena et al 1999; Mataix-Solera et al. 2011).
The ER of small gravel particles > 2000 [micro]m was always less than 1.0 in all treatments (Fig. 3c) and was most probably detached and transported by rainsplash due to low runoff energy at obstructed, small plot scales (5x2 m). In the case of the BUP plots, field evidence suggests that in addition to rainsplash transport, wind and gravity dry ravel mechanisms may have substantially detached and transported small gravelsized particles owing to the higher bare soil exposed in this treatment plots. This is evident with ER values that were closer to 1.0 during some events (Fig. 3c). The results presented in Fig. 3 conform to the common observations that the size distribution of eroded sediments could be affected by aggregate breakdown or stability resistance during erosion, as well as by the soil texture and particle size distribution of the parent soil (Young 1980; Loch and Donnollan 1983; Proffitt and Rose 1991).
The general trend in the relationship between partiele size distribution of the eroded sediments and [I.sub.30] was relatively constant (Fig. 4) due to vegetation cover providing obstacles for sediment detachment and movement on steeper slopes. This result is in contrast with the traditional assumption of a positive relationship between eroded particle size and rainfall runoff flow, which should be only applicable on post-fire hillslopes with smoothed, bare surfaces void of vegetation, residues and woody debris cover. Slattery and Burt (1997) found a decrease in sediment size with increasing discharge at hillslope scale and attributed the outcome to the hydraulic conditions and dynamics of erosion and sediment delivery at various spatial scales. According to Walling and Moorehead (1989), there are situations where eroded particle sizes remains essentially constant, increase or decrease with increasing discharge.
The MWD of the sediments was calculated from the size distribution of the eroded sediments for individual rainfall events and separated for three rainfall intensities as affected by percentage vegetation cover or bare soil exposed (Table 3) and slope (Table 4). One-way ANOVA revealed that eroded sediment MWD in the BUP plots was significantly higher than in the BSP and CP plots following varying rainfall intensities and for all events combined. The MWD in the BUP plots was significantly higher than in the BSP and CP plots, whereas no significant difference was found in eroded sediment MWD between the BSP and CP treatments following varying /30 (Table 3). The results suggest that a significant positive relationship exists between eroded sediment MWD and percentage bare soil exposed, associated with high soil bum severity and soil disaggregation, in the BUP plots. Conversely, a significant negative relationship exists between eroded sediment MWD and percentage vegetation cover, associated with higher vegetation cover and soil aggregation, in BSP and CP plots, independent of rainfall intensity (Tables 3, 4).
The MWD of eroded sediment decreased with rainfall intensity in the BUP plots, whereas no significant relationship was found within the BSP and CP treatments plots following varying [I.sub.30] (Table 3). This indicates that higher MWD is related to small events with lower erosion only in the BUP plots, but these treatment plots showed visible soil disaggregation owing to the higher soil bum severity and percentage of bare soil exposed. This scenario provided favourable conditions for rainsplash transport in the BUP plots during small [I.sub.30] events and for gravity and wind (non-water) transport mechanisms, as suggested by Moody (2010). Overall, the relationship between MWD of eroded sediment and soil aggregation and erosion in the present study does not conform to that found by Mitchell et al. (1983) and Teixeira and Misra (1997), who reported that strongly aggregated soils have higher MWD because of resistance to disaggregation and lower erosion.
In the BSP and CP treatment plots, a significant positive relationship was found between MWD and slope during small events, whereas no significant relationship was found for medium and high events, in line with the findings of Shi et al. (2012). Surprisingly, a significant negative relationship was found between MWD and slope in the BUP plots for all events (Table 4), in contrast with the traditional assumption of an increase in medium to coarser sediment particles (>0.152 mm) with increased slope in excess of 10% (Kinnell 2000; Shi et al. 2012).
Effects of vegetation recovery and slope on sediment transport selectivity and erosion processes
The relationship between sediment ER and maximum [I.sup.30] for the various treatments is shown in Fig. 5. The relationship between maximum [I.sup.30] and sediment ER in the BUP plots was always positively linear ([R.sup.2] = 0.703, P< 0.05; Fig. 5). This indicates increased aggregate breakdown into primary particles that were easily transported owing to the higher percentage bare soil exposed and soil disaggregation. In contrast, very weak or no significant relationship was observed between sediment ER and maximum [I.sub.30], in the BSP ([R.sup.2] = 0.029 P > 0.05) and CP ([R.sup.2] = 0.015, P > 0.05) treatment plots owing to higher ground vegetation cover and strong soil aggregation and structural stability (Fig. 5).
The minimum threshold rainfall intensity necessary to break down aggregate and transport sediments as primary particles in the BUP plots was reached at <3.56 mm [h.sup.-1] and increased until a maximum [I.sub.30] threshold of 10.9 mm [h.sup.-1], whereas it was seldom reached in the BSP and CP treatment plots (Fig. 5) because of >70% vegetation cover. Martinez-Mena et al. (1999) found that the minimum rainfall intensity threshold necessary to break down aggregates into primary particles was not reached in a natural undisturbed plot because of vegetation cover, which provided soil aggregate stability and reduced rainfall erosion energy, unlike in a disturbed plot.
In the CP plots with >75% canopy cover and 100% combined litter, duff and herbs, vines and shrub ground cover, the rainfall kinetic energy was, on average, 17.7-fold lower than that of the BUP plots with 20-30% ground vegetal cover and 5.6-fold lower than that of the BSP plots with 70-80% vegetation and litter cover (Table 5). The average rainfall energy in the BSP plots was also 68% lower than in the BUP plots. The highest [I.sub.30] maximum rainfall intensity of 12.7 [mm.sup.-1] corresponded to kinetic energy values of 17 578, 5547 and 993 J [m.sup.-2] in the BUP, BSP and CP treatment plots, respectively (Table 5).
The runoff energy (stream power;) available for erosion in the CP plots was, on average, 21.3- and 6.3-fold lower than in BUP and BSP plots respectively ([OMEGA] = 0.6, 3.8 and 12.5 J[m.sup.-2] respectively; Table 5). The average runoff energy in the BSP plots was 70% lower than in the BUP plots. Martinez-Mena et al. (1999) found that a ground vegetation cover of 40% in a natural undisturbed plot with 70% canopy cover reduced rainfall kinetic energy by 50% and mean stream power by 4.5-fold than in a disturbed plot. The results suggest that sediment detachment and transport response to rainfall and runoff energy was quicker and higher in the BUP than BSP and CP treatments plots, where the rainfall runoff energy was diminished by a higher percentage of ground vegetation and litter cover providing surface roughness. There is the natural tendency of vegetation cover to dissipate the rainfall and runoff energy available for aggregate breakdown and erosion transport (Proffitt and Rose 1991; Albaladejo et al. 1998; Martinez-Mena et al. 1999).
The relationship between sediment particle size transport selectivity and rainfall intensity as affected by different vegetation cover conditions is shown in Fig. 6. In all percentage vegetation and slope gradient treatment groups, sand-sized particles (63-2000 [micro]m) showed greater enrichment (ER sand values always >1.0) in all events (Fig. 6). However, higher ER values for sand were observed in the BUP plots (Fig. 6a) due to disaggregation of the highly sandy soil following the high soil burn severity. According to Mataix-Solera et al. (2011), unbumed sandy soils that have organic matter as the principal binding agent have higher aggregate stability but experience a significant decrease in aggregate stability and become highly water repellent after a high-severity fire burn due to destruction of the organic matter. In a similar study, Martinez-Mena et al. (1999) found that the ER sand was significantly > 1.0 in both a disturbed and a natural undisturbed plot, along with a positive relationship between ER sand particles and [I.sub.30] in the disturbed plot but no significant differences in ER sand particles in the natural undisturbed plot with varying [I.sub.30]
Sand enrichment was higher and increased with increases in [I.sub.30] in the BUP plots due to interrill transport on bare soil patches compared with the BSP plots with steeper slope gradients of up to 50% (Fig. 6a, b). In general, the results reflect sand sediment availability and supply and transport selectivity dominated by the 125-205 [micro]m size class because of the sand soil texture in all treatment plots. Soil texture is the main factor affecting differences in eroded sediment size distribution and it is concluded that sediment transport of a particular particle size is determined by the sediment size class availability and supply (Young 1980; Loch and Donnollan 1983; Proffitt and Rose 1991)."
In the BUP plots, sand (63-2000 [micro]m) and clay silt-sized (< 63 [micro]m) particles showed appreciable enrichment with ER clay silt values always >1.0 with varying [I.sub.30] (Fig. 6a), reflecting finer particle size transport selectivity associated with poor topsoil aggregate stability in the BUP plots. Gravel-sized particles decreased with varying [I.sub.30] in the BUP plots (Fig. 6a), suggesting the contribution of wind and gravity (dry ravel) mechanisms in the detachment and transport of disaggregated, loose and single-grain particles >2000 [micro]m in the high percentage of bare soil exposed plots. In the BSP and CP treatment plots, the ERs of clay silt and sand-sized particles did not increase significantly with [I.sub.30]. The ER clay silt was mostly less than or closer to 1.0 in the BSP plots, despite their steeper slope gradient of up 50%, and always significantly < 1.0 in the CP plots, reflecting finer particle depiction associated with topsoil aggregate stability in both treatment plots.
The enrichment of particle sizes of <63 [micro]m in the BUP plots with varying [I.sub.30] conforms to the conclusion that there is a tendency for smaller particles to outrun larger particles so that the average size of eroded sediment decreases with event magnitude on individual small plots (Parsons and Stromberg 1998). The higher enrichment of sand- and clay silt-sized particles in the BUP plots (ER>1.0) indicated aggregate breakdown and increased detachment and transport rate of finer and loose single-grained particles with higher raindrop impact. Gabet and Dunne (2003) reported that a good relationship exists between rain power and the effective detachment rate of fine-grained particles by raindrop impact. On disturbed slopes, finer sediments between 20 and 200 [micro]m in size are the most erodible, even under conditions of limited availability, because they do not have enough mass to limit their movement (Young 1980), particularly when rainfall energy is high enough to break soil aggregates and make finer sediments available for transport (Dumford and King 1993). Hairsine et al. (1999) suggest that the enrichment of particle sizes indicates structural aggregate breakdown by raindrop impact and/or particle size transport selectivity because eroded sediment has a degree of aggregation equal to or less than that of the in situ soil; no breakdown in the soil structure at steady state rainfall-driven erosion means that the particle size distribution of the eroded soil will be the same as that of the original soil.
The non-enrichment clay silt-sized particles <63 [micro]m in the BSP (ER [less than or equal to] 1.0) and CP (ER <1.0) treatment plots indicate structural aggregate stability- and detachment and transport-limited conditions sustained by higher vegetation cover in both treatment plots. The presence of fine colloidal materials and organic matter in soils effectively covered by vegetation or undergoing effective vegetation restoration is usually associated with aggregation by cementation and flocculation to form stable aggregates (Bronick and Lai 2005; Mataix-Solera et al. 2011).
The ER values of gravel-sized particles (>2000 [micro]m) were always significantly <1.0 in all treatment plots (Fig. 6) because of lower rainfall runoff energy at small plot scale (Boix-Fayos et al. 2007), unlike on steep, unobstructed disturbed slopes with patchy vegetation cover. In the BUP plots, ER gravel values were higher and showed a negative relationship with [I.sub.30] (Fig. 6a). The relationship between ER gravel and [I.sub.30] was not linear in the BSP and CP plots (Fig. 6b, c).
Multivariate regression analysis between individual particle size distribution and [I.sub.30], runoff flow, percentage of bare soil exposed, slope gradient, bulk density and organic matter for all treatment groups combined is presented in Table 6. The results show that significant linear relationships exist between two or more of the independent variables and particle size classes for all treatment plots combined (Table 6).
Therefore, the observed differences in the eroded sediment ERs, M WD, proportion and the dissimilar pattern of particle size distribution in relation to rainfall intensity between the three treatment plots with varying vegetation cover provide supporting evidence of the effect of vegetation cover on the size distribution of eroded sediment. The different relationships found between particle size classes and percentage of bare soil exposed, [I.sub.30], runoff flow and slope highlight the different conditions limiting sediment detachment and particle size transport selectivity on post-fire hillslope plots with different vegetation conditions. The lack of sediment loss and runoff plot data 1-5 years after the fire hindered any effort to characterise and quantify the changes in erosion conditions that may have affected eroded sediment size distribution and preferential particle size transport immediately after the fire and over time.
The median eroded sediment was coarser than the median uneroded sediment in the BUP plots during all erosive events because much of the eroded sediment was transported as macroaggregates rather than micro aggregates, as was the case in the BSP plots (75%) and CP plots (100%). The eroded sediment particle sizes did not increase with increasing slope and rainfall intensity in the BSP plots with steeper slope and CP plots because of effective vegetation recovery, which reduced rainfall energy by 5.6- and 17.7-fold respectively and runoff energy by 6.3- and 21.3-fold respectively, limiting aggregate breakdown and sediment detachment and transport selectivity of finer clay silt particles <63 [micro]m.
In the BUP plots, the higher percentage of bare soil exposed ensured favourable conditions for sediment detachment and transport by both rainsplash and rainflow mechanisms at a minimum threshold [I.sub.30] of < 3.56 mm [h.sup.-1] and increased with higher rainfall intensity. This was reflected in the greater enrichment and transport selectivity of finer particles < 63 [micro]m and sand-sized particles in the 63-2000 [micro]m size class with higher rainfall intensity. In contrast, the lower percentage of bare soils exposed in the BSP and CP treatment plots limited sediment detachment and transport processes only to rainsplash, independent of rainfall intensity and slope, reflected in the nonenrichment of finer particles < 63 [micro]m. The higher MWD of the eroded sediment was related to higher percentage of bare soil exposed associated with high soil burn severity, soil disaggregation and higher proportion of gravel content in the surface soil, in the BUP plots.
The general results indicate that percentage of bare soil exposed or, conversely, percentage vegetation cover have more substantial effects on the differences in the MWD of the eroded sediment and in the ER of particle sizes between the treatment plots than slope with varying [I.sub.30]. The enrichment of sand in all treatments indicated sand sediment availability and transport selectivity owing to the dominantly sandy soil texture of the study plots and suggests that the sediment sand in the 125-250-[micro]m size class is a suitable indicator of susceptibility to erosion in the study area.
The authors thank Ki Whan Lee, Ju Gyeong Jung, Young Suh Yoon and Jun Ho Hwang for their contribution to collection of field and laboratory data.
Albaladejo J, Martinez-Mena M. Roldan A, Castillo VM (1998) Soil degradation and desertification induced by vegetation removal in a semiarid environment. Soil Use and Management 14, 15. doi: 10.1111/j. 1475-2743.1998.tb00602.x
Badia-Villas D, Gonzalez-Perez JA, Aznar JM, Arjona-Gracia B, Marti-Dalmau C (2014) Changes in water repellency, aggregation and organic matter of a mollic horizon burned in laboratory: soil depth affected by fire. Geoderma 213, 400 407. doi:10.1016/j.geoderma.2013.08.038
Bagnold RA (1966) An approach to the sediment transport problem from general physics. U.S. Geological Survey Professional Paper 422-1. Washington DC.
Benavides-Solorio J, MacDonald LH (2005) Measurement and prediction of post-fire erosion at the hillslope scale, Colorado Front Range. International Journal of Wildland Fire 14, 457 474. doi: 10.1071/ WF05042
Boix-Fayos C, Martinez-Mena M, Calvo-Cases A, Amau-Rosalen E, Albaladejo J, Castillo V (2007) Causes and underlying processes of measurement variability in field erosion plots in Mediterranean conditions. Earth Surface Processes and Landforms 32, 85 101. doi: 10.1002/esp. 1382
Bronick CJ, Lai R (2005) Soil structure and management, a review. Geoderma 124, 3-22. doi:10.1016/j.geoderma.2004.03.005
Cerda A (1998) Changes in overland flow and infiltration after a rangeland fire in a Mediterranean scrubland. Hydrological Processes 12, 1031-1042.
Cerda A, Doerr SH (2005) Influence of vegetation recovery on soil hydrology and erodibility following fire: an 11-year investigation. International Journal of Wildland Fire 14, 423 437. doi: 10.1071/ WF05044
Choung Y, Lee B, Cho J, Lee K, Jang 1, Kim S, Hong S, Jung H, Choung H (2004) Forest responses to the large-scale east coast fires in Korea. Ecological Research 19, 43-54. doi: 10.1111/j.1440-1703.2003. 00607.x
Durnford D, King JP (1993) Experimental study of processes and particle-size distributions of eroded soil. Journal of Irrigation and Drainage Engineering 119, 383 398. doi: 10.1061/(ASCE)0733-9437 (1993)119:2(383)
Farenhorst A, Bryan RB (1995) Particle size distribution of sediment transported by shallow flow. Catena 25, 47-62. doi: 10.1016/0341 8162(94(00041-C
Gabet EJ, Dunne T (2003) Sediment detachment by rain power. WaterResources Research 39, 1 12.
Hairsine PB, Sander GC, Rose CW, Parlange JY, Hogarth WL, Lisle I, Rouhipour H (1999) Unsteady soil erosion due to rainfall impact: a model of sediment sorting on the hillslope. Journal of Hydrology 220, 115-128. doi : 10.1016/S0022-1694(99)00068-2
Heiri O, Lotter AF, Lemcke G (2001) Loss on ignition as a method for estimating organic and carbonate content in sediments: reproducibility and comparability of results. Journal of Paleolinmology 25. 101 110. doi: 10.1023/A: 1008119611481
Kemper WD. Rosenau RC (1986) Aggregate stability and size distribution. In 'Methods of soil analysis, part I'. (Ed. A Klute) pp. 425-442. (American Society of Agronomy: Madison, WI)
Kim CG, Shin KI, Joo KY, Lee KS, Shin SS, Choung YS (2008) Effects of soil conservation measures in a partially vegetated area after forest fires. The Science of the Total Environment 399, 158 164. doi : 10.1016/ j.scitotenv.2008.03.034
Kinnell PIA (2000) The effect of slope length on sediment concentrations associated with side-slope erosion. Soil Science Society of America Journal 64, 1004 1008. doi: 10.2136/sssaj2000.6431004x
Larsen IJ, MacDonald LH, Brown E, Rough D, Welsh MJ, Pietraszek JH, Libohova Z, Benavides-Solorio J, Schaffrath K (2009) Causes of postfire runoff and erosion: roles of soil water repellency, surface cover, and soil sealing? Soil Science Society of America Journal 73, 1393-1407. doi: 10.2136/sssaj2007.0432
Le Bissonnais Y (1996) Aggregate stability and assessment of soil crustability and erodibility: 1. Theory and methodology. European Journal of Soil Science 47, 425 437. doi:10.1111/j.1365-2389.1996. tb01843.x
Lee BD, Park PS, Chung J (2006) Temporal and spatial characteristics of forest fires in South Korea between 1970 and 2003. International Journal of Wildland Fire 15, 389 396. doi:10.1071/WF05090
Lee CW, Seo JI, Youn HJ (2014) Effectiveness of rehabilitation treatments on a slowly revegetating hillslope in a recently burned coastal forest, Republic of Korea. Landscape Ecological Engineering 10, 249-260. doi: 10.1007/s 11355-013-0218-7
Lim JH, Lee YG, Jeong JH, Lee MB (2010) Post-fire restoration to establish a healthy and sustainable forest ecosystem. Publication no. 11-1400377000393-01, Korea Forest Research Institute, Seoul.
Loch RJ, Donnollan TE (1983) Field rainfall simulator studies on two clay soils of the Darling Downs, Queensland. II Aggregate break-down, sediment properties and soil erodibility. Australian Journal of Soil Research 21, 47 58. doi:10.1071/SR9830047
Martinez-Mena M, Rogei JA, Albaladejo J, Castillo VM (1999) Influence of vegetal cover on sediment particle size distribution in natural rainfall conditions in a semiarid environment. Catena 38, 175 190. doi: 10.1016/ S0341-8162(99)00073-9
Mataix-Solera J, Cerda A, Arcencgui V, Jordan A, Zavala LM (2011) Fire effects on soil aggregation: a review. Earth-Science Reviews 109,44 60. doi: 10.1016/j.earscirev.2011.08.002
Meyer LD, Line DH, Harmon WC (1992) Size characteristics of sediment from agricultural soils. Journal of Soil and Water Conservation 47, 107 111.
Miller WP, Baharuddin MK (1987) Particle size of interri 11-eroded sediments from highly weathered soils. Soil Science Society of America Journal 51, 1610-1615. doi: 10.2136/sssaj1987.03615995005 100060037x
Mitchell JK, Mostaghimi S, Pound M (1983) Primary particle and aggregate size distribution of eroded soil from sequenced rainfall events. Transactions of the ASAE. American Society of Agricultural Engineers 26, 1773 1777. doi: 10.13031/2013.33841
Moody JA (2010) Plot-scale sediment transport processes on a burned hillslope as a function of particle size. In 'Proceedings of the 9th Pederal Interagency Sedimentation Conference', 27 June-1 July 2010, Las Vegas, NV. pp. 1-12.
Moody JA, Shakesby RA, Robichaud PR, Cannon SH, Martin DA (2013) Current research issues related to post-wildfire runoff and erosion processes. Earth-Science Reviews 122, 10- 37. doi:10.1016/j.earscirev. 2013.03.004
Ongley ED, Bynoe MC, Perciva JB (1981) Physical and geochemical characteristics of suspended solids. Wilton Creek, Ontario. Canadian Journal of Earth Sciences 18, 1365-1379. doi: 10.1139/e81-126
Parsons AJ, Stromberg SGL (1998) Experimental analysis of size and distance of travel of unconstrained particles in interrill flow. Water Resources Research 34, 2377-2381. doi:10.1029/98WR01471
Parsons AJ, Brazier RE, Wainwright .1. Powell DM (2006) Scale relationships in hillslope mnoff and erosion. Earth Surface Processes and Landforms 31. 1384 1393. doi: 10.1002/esp. 1345
Parsons AJ, Robichaud PR, Lewis SA, Napper C, Clark JT (2010) Field guide for mapping post-fire soil bum severity. General Technical Report RMRS-GTR-243, USDA Forest Service, Rocky Mountain Research Station, Fort Collins, CO.
Proffitt APB, Rose CW (1991) Soil erosion processes: I. The relative importance of rainfall detachment and runoff entrainment. Australian Journal of Soil Research 29, 671 683. doi : 10.1071/SR9910671
Prosser I, Williams L (1998) The effect of wildfire on mnoff and erosion in native Eucalyptus forest. Hydrological Processes 12, 251-265. doi: 10.1002/(SICI) 1099-1085(199802)12:2<251::AID-HYP574>3.0. CO;2-4
Shakesby RA (2011) Post-wildfire soil erosion in the Mediterranean: review and future research directions. Earth-Science Reviews 105, 71 100. doi: 10.1016/j.earscirev.2011.01.001
Shakesby RA, Doerr SH (2006) Wildfire as a hydrological and geomorphological agent. Earth-Science Reviews 74. 269-307. doi: 10.1016/j.earscirev.2005.10.006
Shi ZH, Fang NF, Wu FZ, Wang L, Yue BJ, Wu GL (2012) Soil erosion processes and sediment sorting associated with transport mechanisms on steep slopes. Journal of Hydrology 454-455, 123-130. doi: 10.1016/ j.jhydrol.2012.06.004
Shin SS, Park SD, Lee SK (2013) Sediment and hydrological response to vegetation recovery following wildfire on hillslopes and hollow of a small watershed. Journal of Hydrology 499, 154-166. doi: 10.1016/ j.jhydrol.2013.06.048
Slattery MC, Burt TP (1995) Size characteristics of sediment eroded from agricultural soil: dispersed versus non-dispersed, ultimate versus effective, in 'River gcomorphology'. (Ed. EJ Hickin) pp. I -17. (John Wiley and Sons: Chichester)
Slattery MC, Burt TP (1997) Particle size characteristics of suspended sediment in hillslope mnoff and streamflow. Earth Surface Processes and Landforms 22, 705-719. doi: 10.1002/(SICI) 1096-9837(199708) 22:8<705::AID-ESP739>3.0.CO;2-6
Teixeira PC, Misra RK. (1997) Erosion and sediment characteristics of cultivated forest soils as affected by the mechanical stability of aggregates. Catena 30, 119 134. doi : 10.1016/S0341 -8162(97)00023-4
Walling DE, Moorehead PW (1989) The particle size characteristics of fluvial suspended sediment: an overview. Hydrobiologia 176-177, 125 149. doi: 10.1007/BF00026549
Young RA (1980) Characteristics of eroded sediment. Transactions of the ASAE. American Society of Agricultural Engineers 23, 1139 1142. doi: 10.13031/2013.34735
Ewane Basil Ewane (A,B) and Eleon-Ho Lee (A)
(A) Department of Forest Resources, College of Natural Resources, Yeungnam University, 280 Daehak-ro, Gyeongsan-si, Cyeongsangbuk-do 712-749, South Korea.
(B) Corresponding author. Email: email@example.com
Table 1. Mean ([+ or -] s.d.) values of selected physical properties of the soil in the slow vegetation recovery (BUP), fast vegetation recovery (BSP) and unburned control (CP) plots Within columns, values with different letters differ significantly (P < 0.05, Tukey's test). Only results of analyses of soil samples from the 0-5 cm depth are presented because no significant difference was found in the measures of soil properties between the burned and unbumed plots at soil depths of 5-10 cm. WCV, water content, volumetric; WCG, water content, gravimetric; OM, organic matter; BD, bulk density; GC, gravel content Treatment WCV (g [cm.sup.-3]) WCG (g [g.sup.-1]) BUP 0.04 [+ or -] 0.02a 0.03 [+ or -] 0.01a BSP 0.14 [+ or -] 0.03b 0.11 [+ or -] 0.02b CP 0.13 [+ or -] 0.02b 0.11 [+ or -] 0.02b Treatment Porosity (%) OM (%) BUP 38.8 [+ or -] 7.3a 1.72 [+ or -] 0.17a BSP 51.9 [+ or -] 2.5b 2.58 [+ or -] 0.13b CP 53.5 [+ or -] 1.5b 4.44 [+ or -] 1.29c Treatment BD (g [cm.sup.-3]) GC (%) Soil texture BUP 1.62 [+ or -] 0.19a 47.8 [+ or -] 11.2a Sand BSP 1.27 [+ or -] 0.07b 23.5 [+ or -] 6.7b Sand CP 1.23 [+ or -] 0.04b 12.6 [+ or -] 4.1c Loamy sand Table 2. Comparison of the size distribution of uneroded and eroded sediments in the slow vegetation recovery (BLIP), fast vegetation recovery (BSP) and unburned control (CP) plots Data are the meanis.d. Within rows, values with different letters differ significantly (P < 0.05, Tukey's multiple comparison post hoc test). CV, coefficient of variation Unerodcd sediment Size class ([micro]m) BUP BSP <63 6.55 [+ or -] 1.94a 10.72 [+ or -] 2.45b 63-125 8.98 [+ or -] 1.25a 15.46 [+ or -] 3.73b 125-250 18.82 [+ or -] 3.90a 22.18 [+ or -] 2.04ab 250-500 10.56 [+ or -] 3.04a 13.13 [+ or -] 1.86ab 500-1000 13.13 [+ or -] 1.37a 14.02 [+ or -] 3.75a 1000-2000 8.13 [+ or -] 2.41a 5.21 [+ or -] 1.94b >2000 33.83 [+ or -] 8.60a 19.27 [+ or -] 5.47b CV (%) 67 39 Uneroded sediment Eroded sediment Size class ([micro]m) CP CV (%) BUP <63 20.29 [+ or -] 1.72c 56.2 8.19 [+ or -] 3.86a 63-125 15.15 [+ or -] 1.28b 27.7 13.15 [+ or -] 2.28a 125-250 25.23 [+ or -] 0.86b 14.5 29.44 [+ or -] 3.82a 250-500 16.78 [+ or -] 0.66b 23.2 13.13 [+ or -] 4.17a 500-1000 14.62 [+ or -] 1.52a 5.4 13.28 [+ or -] 1.45a 1000-2000 3.18 [+ or -] 0.75b 45.1 7.89 [+ or -] l.87a >2000 4.75 [+ or -] 2.74c 75.4 14.93 [+ or -] 7.44a CV (%) 56 50 Eroded sediment Size class ([micro]m) BSP CP CV (%) <63 11.25 [+ or -] 6.11a 1.79 [+ or -] 1.98b 48.2 63-125 16.33 [+ or -] 4.03a 7.67 [+ or -] 4.26b 26.2 125-250 31.63 [+ or -] 3.95a 28.38 [+ or -] 14.40a 7.6 250-500 16.56 [+ or -] 3.13a 24.05 [+ or -] 12.64b 41.9 500-1000 14.58 [+ or -] 3.22a 17.71 [+ or -] 13.03a 21.7 1000-2000 5.52 [+ or -] 2.05ab 3.87 [+ or -] 3.95b 29.1 >2000 4.13 [+ or -] 2.45b 0.85 [+ or -] 1.33c 102.5 CV (%) 64 93 Table 3. Effects percentage of bare soil exposed (i.e. slow vegetation recovery (BLIP), fast vegetation recovery (BSP) and unburned control (CP) plots) on mean weight diameter (MWD) of sediments for different natural rainfall intensity conditions Within columns, values with different lowercase letters differ significantly (P < 0.05, Tukey's test). Within rows, values with different uppercase letters differ significantly (P < 0.05, Tukey's test). Small events, 30-min rainfall intensity ([I.sub.30]) < 3.56 mm [h.sup.-1]; medium events, [I.sub.30] 6.1-6.9 mm [h.sup.-1]; high events, [I.sub.30] > 7.6 mm [h.sup.-1] Treatment MWD of eroded sediments ([micro]m) Small Medium High All events events events events combined BUP 947Aa 773Ba 638Bb 797a BSP 498Ab 497Ab 485Ab 495b CP 377Ab 429Ab 523Ab 427b Table 4. Effects of slope in the slow vegetation recovery (BLIP), fast vegetation recovery (BSP) and unburned control (CP) plots on mean weight diameter (MWD) of sediments for different natural rainfall intensity conditions Within columns, values with different lowercase letters differ significantly (P < 0.05, Tukey's test). Within rows, values with different uppercase letters differ significantly (P < 0.05, Tukey's test). Small events, 30-min rainfall intensity ([I.sub.30]) <3.56 mm [h.sup.-1]; medium events, [I.sub.30] 6.1-6.9 mm [h.sup.-1]; high events, [I.sub.30] > 7.6 mm [h.sup.-1] Treatment Slope (%) MWD of sediments ([micro]m) Small Medium High All events events events events combined BUP 11 1190Aa 949ABa 738Ba 977a BUP 22 703Ab 598ABb 539Bb 618b CP 11 376Ac 429Ab 523Ab 427c BSP 17 377Ac 406Ab 415Ab 398c BSP 32 518Abc 507Ab 525Ab 515c BSP 46 601Ab 579Ab 515Ab 570c Table 5. Mean values of rainfall properties and erosion responses in the slow vegetation recovery (BLIP), fast vegetation recovery (BSP) and unburned control (CP) plots The rainfall properties evaluated were maximum 30-min rainfall intensity ([I.sub.30]), stream power ([OMEGA]) and rainfall kinetic energy (ICE). The erosion responses evaluated were run-off flow ([Q.sub.w]) and sediment loss ([S.sub.w]). BSE, bare soil exposed (or, conversely, vegetation cover; e.g. 70-80% BSE is equivalent to 20-30% vegetation cover); [I.sub.30], 30-min rainfall intensity. Only events for which data for both eroded and uneroded sediments were collected are presented Event date Rainfall [I.sub.30] (mm) (mm [h.sup.-1]) 20.6.2014 13 2.5 26.6.2015 32 10.9 7.7.2015 29 6.9 9.7.2015 114 12.7 13.7.2015 36 6.9 20.7.2015 5 3.1 12.8.2015 15 3.1 21.8.2015 31 3.6 25.8.2015 49 6.1 1.9.2015 17 7.6 5.9.2015 23 3.6 1.10.2015 29 4.3 Event date Burned plots BUP (70 80% BSE) [Q.sub.w] [S.sub.w] [OMEGA] KE (J (L (g (J [m.sup.-2] [m.sup.-2] [m.sup.-2]) [m.sup.-2]) [s.sup.-1] h) 20.6.2014 405.7 318.8 20.9 1351.2 26.6.2015 289.0 326.0 33.2 4263.8 7.7.2015 167.3 127.5 11.6 2676.1 9.7.2015 497.5 643.0 9.8 17 578.9 13.7.2015 190.4 191.0 7.5 4622.3 20.7.2015 11.5 24.4 4.3 216.3 12.8.2015 45.1 41.4 2.7 1406.1 21.8.2015 175.5 179.5 15.2 1039.7 25.8.2015 173.1 163.0 5.9 4759.2 1.9.2015 106.7 86.1 18.6 1162.0 5.9.2015 29.5 28.3 14.3 185.4 1.10.2015 118.5 134.9 5.6 2298.0 Event date Burned plots BSP (20-30% BSE) [Q.sub.w] [S.sub.w] [OMEGA] KE (J (L (g (J [m.sup.-2] [m.sup.-2] [m.sup.-2]) [m.sup.-2]) [s.sup.-1] h) 20.6.2014 22.3 31.5 2.0 426.4 26.6.2015 33.8 38.2 4.7 1345.6 7.7.2015 40.0 12.9 5.5 844.6 9.7.2015 122.9 97.3 5.3 5547.9 13.7.2015 27.2 19.1 2.1 1458.8 20.7.2015 4.90 3.5 3.8 68.3 12.8.2015 13.4 8.9 1.6 443.8 21.8.2015 20.9 20.7 4.1 328.1 25.8.2015 38.1 24.8 2.6 1502.0 1.9.2015 8.47 6.3 3.0 366.7 5.9.2015 7.63 3.3 8.1 58.5 1.10.2015 31.0 10.6 3.1 725.3 Event date CP (0% BSE) [Q.sub.w] [S.sub.w] [OMEGA] KE (J (L (g (J [m.sup.-2] [m.sup.-2] [m.sup.-2]) [m.sup.-2]) [s.sup.-1] h) 20.6.2014 13.7 0.7 0.5 76.3 26.6.2015 11.92 0.7 0.6 240.9 7.7.2015 14.5 0.5 0.7 151.2 9.7.2015 25.9 4.6 0.4 993.1 13.7.2015 8.38 7.3 0.3 261.1 20.7.2015 2.58 0.7 0.7 12.2 12.8.2015 5.30 0.0 0.2 79.4 21.8.2015 4.59 0.8 0.3 58.7 25.8.2015 13.7 0.5 0.3 268.9 1.9.2015 3.34 0.9 0.4 65.6 5.9.2015 4.26 0.7 1.6 10.5 1.10.2015 14.0 2.8 0.5 129.8 Table 6. Multiple linear regression equations relating individual particle size distribution (PSD) with rainfall intensity (30-min rainfall intensity; ho)i runoff How (Q), percentage of bare soil exposed (BSE), slope gradient (S), bulk density (BD) and organic matter (OM) for all treatments combined and for individual slow vegetation recovery (BUP), fast vegetation recovery (BSP) and unburned control (CP) plots * P < 0.05. [R.sup.2], coefficient of determination; s.e.e., standard error of estimate No. size Equation limits Size class ([micro]m) <63 PSD = [I.sub.30] * + Q * + S * + BD * + BSE * + OM * 63-125 PSD = [I.sub.30] * + Q * + S * + BD * + BSE * + OM 125-250 PSD = [I.sub.30] + Q + S + BD + BSE + OM 250-500 PSD = [I.sub.30] * + Q + S + BD + BSE + OM * 500-1000 PSD = [I.sub.30] * + Q + S + BD + BSE * + OM 1000 2000 PSD = [I.sub.30] + Q + S + BD + BSE + OM >2000 PSD = [I.sub.30] + Q + S + BD + BSE + OM Treatment BUP 7 PSD = Q + [I.sub.30] + S i BSE BSP 7 PSD = Q + BSE CP 7 PSD = Q + S [R.sup.2] Adjusted [R.sup.2] s.e.e. P-value Size class ([micro]m) <63 0.935 0.869 0.003 * 63-125 0.95 0.9 0.001 * 125-250 0.216 0.569 0.929 250-500 0.876 0.753 0.016 * 500-1000 0.64 0.28 0.251 1000 2000 0.613 0.227 0.295 >2000 0.961 0.523 0.092 Treatment P-value BUP 0.909 0.819 3.058 0.045 BSP 0.471 0.365 7.058 0.05 CP 0.551 0.42 1.023 0.028
Please note: Some tables or figures were omitted from this article.
|Printer friendly Cite/link Email Feedback|
|Author:||Ewane, Ewane Basil; Lee, Heon-Ho|
|Date:||Nov 1, 2016|
|Previous Article:||Mineralogy of volcanically derived alluvial soils at Moshi, Tanzania.|
|Next Article:||An alternative index to the exchangeable sodium percentage for an explanation of dispersion occurring in soils.|