| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a USDA-ARS, Northwest Watershed Research Center, 800 Park Boulevard, Plaza IV, Suite 105, Boise, ID 83712
b Department of Plant and Soil Science, Texas Tech University, Lubbock, TX 79409
c Department of Range, Wildlife and Fisheries Management, Texas Tech University, Lubbock, TX 79409
* Corresponding author (cmoffet{at}nwrc.ars.usda.gov)
Received for publication March 1, 2004.
 |
ABSTRACT |
|---|
 TOPNOTES ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Abbreviations: ANOVA, analysis of variance • OC, organic carbon
|
INTRODUCTION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Application of biosolids has ecological effects, both chemical and physical in nature (Epstein, 1975; Gupta et al., 1977; Kladivko and Nelson, 1979; Khaleel et al., 1981; Pagliai et al., 1981; Glauser et al., 1988; Fresquez et al., 1990; Aguilar and Loftin, 1992, 1994; Martens and Frankenberger, 1992; Bruggeman and Mostaghimi, 1993; Pierzynski, 1994; Harris-Pierce et al., 1995; Benton and Wester, 1998; Sort and Alcaniz, 1999; Rostagno and Sosebee, 2001a, 2001b; Wester et al., 2003). Among these ecological effects, hydrological processes including infiltration and erosion are of particular interest in the arid and semiarid western United States where biosolids are typically surface-applied rather than incorporated into the soil.
Studies that have investigated biosolids effects on vegetation generally focus on biosolids features that have obvious meaning in the context of plant growth. For example, biosolids are composed of both mineral and organic matter. Typically, biosolids are 50 to 60% organic matter when fresh (Rostagno and Sosebee, 2001a, 2001b). Biosolids also are a significant source of plant nutrients, notably N, Fe, and P. However, the organic fraction of biosolids has a number of characteristics that are important hydrologically. For example, the organic fraction can be divided into subfractions including fats, waxes, and oils; polysaccharides; hemicellulose; cellulose; lignin-humus; and protein (Smith and Peterson, 1982). Greases, oils, and fats account for 5 to 20% of the total solids (Smith and Peterson, 1982; Outwater, 1994, p. 8–9). As these compounds leach out of the biosolids and onto the soil, they might have important hydrological consequences such as lower clay dispersibility and partial wettability.
Aguilar and Loftin (1992)(1994) and Harris-Pierce et al. (1995) studied infiltration, erosion, and surface water quality as affected by biosolids in semiarid blue grama [Bouteloua gracilis (Willd. ex Kunth) Lag. ex Griffiths] grasslands. Harris-Pierce et al. (1995), from simulated rainfall, reported no effect of biosolids application (22 and 41 dry Mg ha–1) on runoff to precipitation ratios, but reported increased steady-state infiltration rate and sediment content in runoff with increased biosolids application rate. Aguilar and Loftin (1992)(1994) studied runoff and erosion under natural and simulated rainfalls. Six natural storms produced runoff during the study, four of which had significantly less runoff from biosolids-amended treatments (45 dry Mg ha–1). In simulated rainfall trials, runoff was generally significantly less in treated plots. In both studies, decreased runoff was attributed to increased surface roughness due to the biosolids. Additionally, Aguilar and Loftin (1992)(1994) attributed the reduced runoff to absorption of water by the biosolids.
In a companion study, Rostagno and Sosebee (2001a)(2001b) examined the effects of surface-applied biosolids rate (0, 7, 18, 34, and 90 dry Mg ha–1) and post-application age on runoff water quality and soil physical properties of two soil types in the Chihuahuan Desert. They reported that the concentrations of most of the compounds measured were strongly affected by soil type, post-application age, rate, and the interaction between rate and post-application age (Rostagno and Sosebee, 2001a). Rostagno and Sosebee (2001b) reported that surface-applied biosolids have the potential to increase the time-to-runoff and steady-state infiltration rate of degraded soils. They also reported that biosolids application decreased the clay dispersibility and increased organic matter in the upper 3 cm of soil. These authors reported that "the mechanism by which surface-applied biosolids increase soil IR [infiltration rate] is not totally clear," but suggested that their observations supported the increased surface roughness mechanism that Aguilar and Loftin (1992)(1994) suggested as the major factor.
Meyer et al. (2001) studied the effect of incorporated biosolids applications (0, 40, and 80 dry Mg ha–1) two years post-application in a forested site that wildfire had burned three years prior. The site, at the time of rainfall simulation, was primarily a thickspike wheatgrass [Elymus macrourus (Turcz.) Tzvelev]–streambank wheatgrass [Elymus lanceolatus (Scribn. & J.G. Sm.) Gould subsp. lanceolatus] community. The 12 plots used in the study were 30 m2. The soils of their study were gravelly clay-loam to gravelly sandy-loam on slopes ranging from 10 to 16%. The effect of biosolids on runoff was insignificant; however, sediment concentrations were significantly reduced by biosolids application rate. Meyer et al. (2001) attributed this significant reduction in sediment concentration to greater canopy cover found in the biosolids-treated plots.
The objectives of our research were to: (i) evaluate the effects of four rates of surface-applied biosolids on infiltration and erosion on two different soils, with different degrees of canopy cover, measured from simulated rainfall in different seasons, and at different post-application biosolids ages; and (ii) examine some of the ways biosolids may be affecting infiltration and erosion. We tested the hypotheses that infiltration and erosion were not affected by biosolids application rate, biosolids age, season of biosolids application, season of rainfall simulation, or soil type (site and cover). In general, we hypothesized that infiltration and steady-state infiltration rate would increase and erosion decrease with increasing biosolids application rate and biosolids age. We did not expect season of application or simulation to have an effect on infiltration or erosion. We hypothesized that infiltration and steady-state infiltration rate would be greater in the vegetated soil than in bare soils. These hypotheses were tested with analyses of variance.
|
MATERIALS AND METHODS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
The soil parent material was locally derived, mostly igneous, alluvium. The two sites studied were Stellar and Chilicotal. The Stellar site was on Pleistocene-age alluvial flat geomorphic surfaces. The surface was relatively smooth on the Stellar site, with many 5-cm-tall raised surface elements (tussocks), 0.5 to 4 m in diameter, under grass vegetation separated by nearly level and featureless bare areas of lateral extent approximately equal to the tussocks. The Chilicotal site was on Holocene-age inset fan geomorphic surfaces. The surface was sharply dissected by small gullies up to 50 cm deep and 1 to 2 m wide with narrow 20-m-wide interfluves.
Given soil names are field names only and are not representative of the series for which they are named. Hudspeth and adjacent Culberson Counties do not have published soil surveys and series taxa suitable for these soils have not been established. Chilicotal very fine sandy loam (fine-loamy, mixed, thermic Ustic Calciargids) was a taxadjunct of the Chilicotal series. Stellar very fine sandy loam (fine, mixed, thermic Vertic Paleargids) was a taxadjunct of the Stellar series. Slopes were 1% and 3 to 6% for the Stellar and Chilicotal sites, respectively. Soil crusts were well formed on the bare microsites for both soils. Typically, the vegetated microsites on the Stellar soil were not crusted. On the Stellar site there were significant soil texture, structure, and organic carbon (OC) content differences between vegetated and bare microsites. The differences in soil between vegetated and bare areas were not taxonomically important, but were believed to be hydrologically important.
Vegetation was significantly different between sites. The Chilicotal site was a shrubland dominated by creosotebush [Larrea tridentata (Sessé & Moc. ex DC.) Coville] and mesquite (Prosopis glandulosa Torr.), with a minor herbaceous component of mostly fluff grass [Dasyochloa pulchella (Kunth) Willd. ex Rydb.] in the shrub interspace, and bush muhly (Muhlenbergia porteri Scribn. ex Beal) at shrub bases. Only bare shrub interspaces were studied on the Chilicotal site. The Stellar site had a significant herbaceous component of dominantly tobosagrass (Pleuraphis mutica Buckl.) and alkali sacaton [Sporobolus airoides (Torr.) Torr.], with widely scattered mesquite. Additional descriptions of the study site may be found in Benton and Wester (1998) and Rostagno and Sosebee (2001a)( 2001b). Neither site was grazed by domestic livestock for several years before this study.
Biosolids
Four batches of biosolids were used in this study. These biosolids were obtained from a commercial applicator contracting with the city of New York, NY, to land-apply their residential biosolids. Each batch was derived from a separate shipping container about every six months during the study. Five biosolids samples were collected from each batch, frozen, and shipped to the Soil, Water, and Air Testing Laboratory at New Mexico State University, Las Cruces, for chemical analysis. A 10-g subsample was taken from each sample to determine water content before freezing. The water content was determined gravimetrically by heating in a microwave oven and re-weighing at 1-min intervals until the mass stabilized. The biosolids were black and had a greasy consistency when fresh. After the biosolids were surface-applied, they began to dry and shrink, becoming hard masses.
Experimental Design
Infiltration and erosion measurements were made from 380 plots. The general experimental design for these experiments was a split-split-splitplot arrangement of a completely randomized design with repeated measures. The main-plot effect was site (Stellar or Chilicotal). The subplot effect was a factorial combination of the post-application age (time between application and rainfall simulation; 1, 6, or 12 mo) and season (season when rainfall simulations were performed; summer or winter; only summer for the Chilicotal site). The sub-subplot factor was the cover effect (vegetated or bare; only bare in the Chilicotal site). The sub-sub-subplot factor was biosolids application rate (0, 7, 34, and 90 dry Mg ha–1), which correspond to the control, a low commercially viable rate, a high commercially viable rate, and an experimental noncommercial rate. The highest rate (90 dry Mg ha–1) was selected to determine if there were negative effects from a rate that was significantly greater than recommended. Collection period during the rainfall simulation (2.5, 5, 7.5, 10, 15, 20, 25, and 30 min) was the repeated measures effect. There were five replications. Table 1 indicates the relationship among treatments down to the level of the sub-sub-subplot factor (biosolids application rate). Analyses of variance (ANOVA) were performed on balanced subsets of the data to address specific questions (e.g., hypotheses involving site effects were tested with a subset that includes two sites, three ages, in winter, with bare cover treatment, four rates, and eight collection periods only, as the Chilicotal site data set included only one level of season and cover effects). Mean separations were performed by the least significant difference (LSD) test only if the effect was significant in the ANOVA. The LSD was computed using the appropriate standard error and weighted tabular t values as discussed in Gomez and Gomez (1984).
|
Rainfall Simulation
A portable single-nozzle rainfall simulator, similar to the ones used by Elkins (1983), Wilcox et al. (1986), and Spaeth (unpublished data, 1990), was used to simulate the rainfall. The nozzle was placed 2 m above the soil surface, and a tarp was placed around the simulator to protect the simulated rainfall from wind disturbance. A nozzle pressure of 20.7 kPa was maintained during simulation to produce a mean rainfall intensity of 164 mm h–1 on a square 0.5-m2 plot. A storm of this intensity is likely to occur for 10- to 30-min durations every 2 to 3 yr in this region. Median drop size was between 1.2 and 2 mm with the largest drops concentrated in the center of the plot (Wilcox et al., 1986).
A 1.5-mm-diameter water drop falling a distance of 2 m in still air has an impact velocity of 4.50 m s–1, which is 82% of terminal velocity (Bubenzer, 1979). Because the water at the nozzle orifice in this simulator was under pressure, it is reasonable to assume that the impact velocity was greater than that of free-falling drops.
Water from the local municipal water source was used for the simulated rainfall. The water was moderately alkaline with a pH of approximately 8.2. The electrical conductivity was 0.7 dS m–1 and the total dissolved solids were 444 mg L–1. The sodium absorption ratio of the water was 6.7.
A square 0.5-m2 steel frame was positioned on each plot, driven into the ground, and fitted with a runoff collection pan. The runoff water was allowed to flow from the plot into a low corner of the collection pan. The runoff collection pan was sheltered from the simulated rain.
The rainfall simulator was positioned level on each plot and a 30-min rainfall was simulated. Runoff water that collected in the lower corner of the collection pan was transferred through a nylon hose and suction pump to a graduated cylinder. The volume was recorded for each period. Runoff was recorded at eight periods during the simulation, once every 2.5 min for the first 10 min and once every 5 min for the remaining 20 min.
The applied rainfall volume was estimated for each plot. At the end of the 30-min simulation period a calibration pan with the same dimensions as the runoff plot was placed over the plot for a period of 5 min. At the end of the 5-min period, the simulator was shut off and the volume of water collected in the calibration pan was measured. The values for each plot receiving simulated rainfall during a day were averaged, and the average was used as the estimate for all plots measured that day.
After runoff volume was determined for each collection period, runoff water was poured into a clean bucket. At the end of each of three 10-min periods, the runoff water collected during the period was thoroughly mixed and a subsample (approximately 1 L) was collected. The subsample volume was measured precisely in the laboratory and the sample was filtered to remove the sediment. The sediment was oven-dried at 105°C for 24 h and weighed to determine the sediment concentration for that period. The sediment concentration was multiplied by the runoff volume during that period to determine the quantity of erosion.
Completion times varied for each of the three trials (winter 1994, summer 1994, and winter 1995). Effort was made to complete each trial in a short time period. Completion times for the three trials required between 47 and 66 d. Therefore, the actual post-application age of the biosolids at the time of simulation varies (±1 mo).
Plot Characterization
After rainfall simulation was performed, soil microtopography and cover conditions in each plot were characterized using methods similar to Weltz et al. (1992). An ocular estimate of canopy cover was made immediately following the rainfall simulation. Canopy cover was measured in several plots by lowering point frame pins through the canopy and recording the number of contacts with vegetation for 81 points in a grid over the plot (9 rows by 9 columns) in preliminary work to calibrate the ocular estimate. The standing crop mass was then clipped as low as was possible without disturbing the biosolids in the plot. A leveled platform was placed around each plot border and a point frame was attached to the platform along nine (7.5-cm spacing) nine-point transects (7.5-cm spacing). The surface hit was classified and the relative elevation (the length of pin above a datum on the point frame) was measured to the nearest millimeter. If the hit was classified as biosolids, the biosolids were then removed and the point below all pieces of biosolids was classified and its relative elevation measured as above. From these data the volume of depressional storage, maximum relief, mean relief, random roughness, and fractional coverage by classes of ground cover were determined. Cover classes included biosolids, litter, living plant base, dead plant base, microphyte, and pebble (2–75 mm); bare ground was classified into two classes: crust or open crack (between crust polygons).
Within a week after the rainfall simulation was performed soil samples were collected from the crust, A horizon, and B horizon for determination of soil particle size distribution by the hydrometer method (Gee and Bauder, 1986), and OC by the method of Walkley and Black (Nelson and Sommers, 1982). Additional samples were taken from the A and B horizons for determination of bulk density by the coated clod method (USDA, 1992). Water repellency was observed in some plots during sampling. The repellency was qualitatively classified according to the time required for a bead of water to wet the surface similar to the water drop penetration time methods described by Dekker and Ritsema (1994).
|
RESULTS AND DISCUSSION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
= 0.10, but not at the = 0.05 level). The effect of biosolids application rate on OC content did not interact with soil and cover treatments. Application of 7 dry Mg ha–1, or more, of biosolids significantly increased OC content in crusts above that measured in nontreated plots. Further, significant increases in crust OC content were observed at the 34 and 90 dry Mg ha–1 biosolids application rates. Rostagno and Sosebee (2001b) sampled a thinner sub-crust layer (from 0.5–3 cm) than in the present study and reported significantly more organic matter in plots treated with biosolids.
|
= 0.05). The crust OC regression equation:
 | [1] |
Greater depth resolution would have allowed for a better assessment of the impact of biosolids on OC content below the crust. Kladivko and Nelson (1979) reported greater increases in OC in the upper 5 cm of a Celina soil (fine, mixed, mesic Aquic Hapludalfs) in the first 6 mo after surface application of 56 Mg ha–1 of biosolids. The Stellar and Chilicotal soils, however, did not show a strong increase in OC content with increased biosolids rate in the A horizon, but these soils were sampled by horizon and on average the A horizon was significantly thicker (see Table 2) than the 5 cm of soil sampled by Kladivko and Nelson (1979); also, mean annual precipitation is significantly less at this study site than at Wanatah, Indiana, the study site of Kladivko and Nelson (1979).
Much of the organic fraction of biosolids is composed of greases, oils, and fats (Smith and Peterson, 1982; Outwater, 1994, p. 8–9) that can move with infiltrating water into the soil. Runoff water had a distinct yellow hue indicating dissolved organic material was present. The crusts tended to be slightly water repellent (as in Dekker and Ritsema, 1994) especially directly beneath biosolids aggregates (i.e., water drops remained on the soil surface for 3 to 10 s versus infiltrating immediately in nontreated plots and most areas not beneath biosolids aggregates in treated plots). Guidi et al. (1983) reported that soil stability was increased more with fats, waxes, oil, and resins extracted from biosolids than with water-soluble polysaccharides.
The OC content of the vegetated cover treatment differed from the bare treatment in the crust, but not in the A horizon. This is due to the different environmental conditions between the very near surface in the bare and vegetated sites and its effect on organic matter additions due to root production and mortality. Crusts in the bare treatments were subject to dramatic diurnal temperature fluctuations that probably limited root activity in the upper 0.5 cm of soil, whereas in the vegetated treatments litter buildup and shade from the canopy makes the near surface environment more hospitable for root growth. Hahm and Wester (2004) reported maximum summer soil temperatures for bare soils at a depth of 0.6 cm commonly in excess of 50°C at a nearby location. Desert plant roots near the soil surface can be injured by these high temperatures (Nobel, 1997); consequently, it is unlikely that there has been significant OC addition from roots in the near surface of bare plots.
Biosolids Effect on Canopy and Ground Cover
As expected, ground cover and average biosolids thickness increased with increased biosolids application rates (Table 3). On bare Stellar and Chilicotal soil, biosolids depth is linearly related to biosolids application rate:
 | [2] |
|
 | [3] |
This equation may not be valid for rates greater than 90 dry Mg ha–1.
The 90 dry Mg ha–1 biosolids application rate provided an average of 80% biosolids cover in bare plots and about 60% biosolids cover in the vegetated plots. The mean biosolids depths for the 90 dry Mg ha–1 rate in the bare and vegetated treatments were 20 and 27 mm, respectively. The biosolids tend to be suspended in the lower grass in the vegetated plots that causes apparently thicker and less uniform cover of biosolids. Harris-Pierce et al. (1995) observed poor biosolids–soil surface contact because of vegetation on a Colorado grassland.
We did not expect the variation in biosolids cover observed within a rate of application among the different batches of biosolids. For a given soil and cover combination the application of Batches 2 and 4 consistently resulted in a more uniform and thicker biosolids cover per unit of biosolids applied than the application of Batches 1 and 3 (Table 3). Batch 1b applied to the Chilicotal site in the summer of 1993 was more similar to Batches 2 and 4 than to Batches 1 and 3 in terms of water content at the time of application and the amount of cover provided. Typically, for 90 dry Mg ha–1 application rates on bare soils, the difference in cover between odd- (not including Batch 1b) and even-numbered batches is approximately 18%. The odd-numbered batches correspond to summer-season applications (typically drier biosolids at the time of application, see Table 4) and even-numbered batches correspond to winter-season applications (typically wetter biosolids at the time of application, see Table 4).
|
Canopy cover in the bare treatments was not affected by biosolids application. In the vegetated Stellar soil canopy covers were similar (62%) among the 0 to 34 dry Mg ha–1 biosolids application rates. Canopy cover for the 90 dry Mg ha–1 rates was less (49%) compared with other rates due to burial of the canopy under the biosolids. Benton and Wester (1998) reported that season of application was an important factor influencing plant response during the first growing season after application. They proposed that greater productivity from dormant (winter) season application was a result of greater residence time that increased available nutrients and possibly reduced evaporative loss during spring because of a mulching effect. A greater plant response could have hydrological implications due to increased canopy cover (Meyer et al., 2001). The difference in plant response between winter and summer biosolids applications did not result in differences in canopy cover in the vegetated plots. This is due to the significant amount of dead standing cover that was present in the plots before biosolids application, which masked differences in new growth cover.
Infiltration Response to Biosolids Application
Infiltration increased due to biosolids application. In general, the greatest increase in infiltration due to biosolids application occurred in soils where the infiltration rate was low to start with.
Infiltration rates were measured in Stellar soil for both vegetated and bare conditions. There were notable differences in soil properties between these two cover conditions that affect infiltration. Soil textures in the crust (upper 0.5 cm) and A horizon were significantly different ( = 0.05) between the two cover conditions (Table 5). Bare soil, unprotected from raindrop impact, was crusted and had a lower steady-state infiltration rate than the vegetated soil regardless of biosolids application rate (Fig. 1)
. A raindrop impact crust forms when raindrops strike the bare surface and disaggregate the soil near the surface allowing silt and clay to fill and plug soil voids. The sand grains that remain on the surface are free to move by wind once the surface dries. Much of this loose sand will be moved and deposited in coppice dunes beneath nearby plant canopies. The surface of the vegetated soil was typically several centimeters higher than the surrounding bare soil. The sum of the crust and A-horizon thickness was 2.2 cm greater for the vegetated than for the bare soil. The surface soil in the vegetated area was generally thicker and sandier, had a greater number of roots and pores, and did not have the raindrop impact crusts found in the bare areas.
|
|
0.05, Table 6). Among these significant effects, 77% of the total variation was accounted for with cover (21.3%), rate (7.8%), and time into the simulated rainfall (43.8%) main effects, and their interactions (4.5%). The infiltration rate means (averaged across age and season) for the three-way interaction (cover by biosolids rate by time) are plotted in Fig. 1.
|
Bare plots had significantly lower steady-state infiltration rates than vegetated plots ( = 0.05; Table 6, Fig. 1). These results occurred regardless of the biosolids application rate. High biosolids application rates (34 and 90 dry Mg ha–1) extended the duration of preponded- and transient-infiltration and elevated steady-state infiltration rate (Fig. 1). Rostagno and Sosebee (2001b) similarly reported increased time-to-runoff in plots treated with biosolids. In the bare treatment, the application of 90 dry Mg ha–1 of biosolids increased the period of transient-infiltration by about 12 min beyond that of the 0 dry Mg ha–1 treatment (Fig. 1A). On bare 90 dry Mg ha–1 plots pre-ponded infiltration occurred for the first 2.5 to 5 min, but on 0 dry Mg ha–1 plots pre-ponded infiltration occurred for less than 2.5 min. Steady-state infiltration on the bare 0 dry Mg ha–1 plots began at around 7.5 min into the simulated rain whereas on the bare 90 dry Mg ha–1 plots steady-state infiltration did not begin until 20 min into the simulated rain.
The ANOVA for cumulative infiltration from this data set (Table 6) indicated that 77% of the total variation was explained with just 5 of the 15 effects tested. Of the 77% explained, 64% was attributed to the cover condition main effect and an additional 21% was attributed to the rate of biosolids main effect (Table 6). Among the other significant effects were some of the interactions that included age and season, but these effects were small by comparison with the significant main effects. Cumulative infiltration was significantly greater with 90 dry Mg ha–1 than all other biosolids application rates, except for the vegetated treatment in summer in which the 90 dry Mg ha–1 rate did not differ from the 34 dry Mg ha–1 rate (Fig. 2) . Cumulative infiltration was significantly greater in plots treated with 34 dry Mg ha–1 of biosolids than in plots without biosolids applied. The application of 7 dry Mg ha–1 of biosolids, however, resulted in cumulative infiltration that was not significantly different from plots without biosolids applied (Fig. 2).
|
|
= 0.05) than in the A horizon, unlike the crust (0- to 0.5-cm depth) in the vegetated Stellar soil (Table 5). The Chilicotal soil was occupied by significantly more ( = 0.05) volume of coarse material (pebbles and stones) throughout the soil than were the bare and vegetated Stellar soils (Table 5). The steady-state infiltration rate in the nontreated Chilicotal plots was greater than that of the nontreated Stellar plots (Tables 7 and 8). In both soils the infiltration rate curves (infiltration rate at each sample time) for the nontreated plots were very similar among the post-application ages (data not shown) as was expected because the distinction among post-application ages of a nontreated plot is meaningless (i.e., nontreated plots really have the same post-application age treatment whether 1, 6, or 12 mo old). In both soils, infiltration rates at any given time during the simulation were more likely to be different among post-application age treatments at high biosolids application rates than at low biosolids application rates (data not shown).
|
The effects of post-application age differed between soils. For example, in the 90 dry Mg ha–1 biosolids application rate, the steady-state infiltration rates in the 6- and 12-mo-old applications on the Chilicotal soil were the greatest with significantly lower steady-state infiltration rates for the 1-mo old applications; but, in the Stellar soil, the steady-state infiltration rate for the 6-mo old application was significantly greater than both the 1- and 12-mo-old applications (Table 8). Regardless of soil the 6-mo-old post-application age plots had greater steady-state infiltration rates than the 1-mo-old plots, which indicated that as the biosolids age the effect of the biosolids is greater. Twelve-month-old biosolids were equal at increasing steady-state infiltration rates to the 6-mo-old on the Chilicotal soil, but not on the Stellar soil. The pattern of similarity between post-application ages was similar for biosolids depth and biosolids cover data for these summer of 1994 simulated plots. An alternative to post-application age as the important factor for these effects is biosolids batch, which we explore below.
The statistical model for infiltration rate had 15 terms and explained 86% of the total variability (Table 7). Of the 15 effects tested, 13 were significant; however, 73% of the total variation was due to three terms: time into the simulated rainfall (47.8%), rate of biosolids applied (21.9%), and their interaction (2.9%) (Table 7). An additional 6% of the total variation was attributed to post-application age and the post-application age by rate interaction (Table 7).
The effects of different post-application ages and rates of biosolids on cumulative infiltration are given in Table 8. Of seven effects tested in the ANOVA statistical model for cumulative infiltration, six were significant (Table 7). The model explained 76% of the total variation in cumulative infiltration. Biosolids application rate accounted for 49% of the total variation.
For the Chilicotal and bare Stellar soils without application of biosolids, infiltration was 2.1 and 1.7 cm, respectively, regardless of the post-application age (a meaningless factor for the 0 dry Mg ha–1 rate; Table 8). At biosolids application rates of 34 and 90 dry Mg ha–1, infiltration differed among the three ages of biosolids on both Chilicotal and bare Stellar soils (Table 8). For the Chilicotal soil treated with 34 and 90 dry Mg ha–1 of biosolids cumulative infiltration was significantly greater for 6- and 12-mo post-application ages than for the 1-mo post-application age. For the bare Stellar soil treated with 34 and 90 dry Mg ha–1 of biosolids, however, cumulative infiltration was significantly greater for the 6-mo post-application age than for the 1-and 12-mo post-application ages. These results are similar to those reported for steady-state infiltration rate.
Considering only the Chilicotal soil, one might argue that at least at the two higher application rates, the biosolids effects on steady-state infiltration rate increased with increasing biosolids age (Table 8). When, however, the bare Stellar soil is considered it is less clear that the effect of biosolids on steady-state infiltration rate increased with increasing age (Table 8). The answer may lie in the relationship among post-application age, season of application, and biosolids batch.
Factors that were deemed important a priori were enumerated and designed into the study; differences in infiltration and erosion, however, were not expected to vary significantly among plots receiving different batches of biosolids or between seasons of application. The experimental design required that different batches of biosolids were used to implement the season in which rainfall was simulated and post-application biosolids age treatments (i.e., four applications separated in time by about 6 mo, which means there were two winter and two summer applications).
To explore the role of biosolids batch, a data set that included bare and vegetated Stellar soils treated with 90 dry Mg ha–1 of biosolids from three rainfall simulation trials (winter 1994, summer 1994, and winter 1995) and all available post-application ages was analyzed. This data set included two trials/post-application ages with biosolids Batch 1, three trials/post-application ages with Batch 2, two trials/post-application ages with Batch 3, and one trial/post-application age with Batch 4 (Fig. 3) .
|
0.05, 70.7, and 73.9% wet-weight basis for summer and winter, respectively). When you take into account the effect of biosolids batch (e.g., by comparing different post-application ages from plots treated with the same batch of biosolids) there is little evidence that biosolids post-application age has an effect on infiltration. Aguilar and Loftin (1992)(1994) suggested that the difference in runoff between biosolids-treated and nontreated plots could be explained by increased surface roughness and interception loss. At the time of application, the biosolids used in this study contained between 2.8 and 3.6 cm of water. Biosolids volume is significantly reduced by the loss of water in the first several days post-application. After drying, biosolids do not swell to accommodate the absorption of water during rewetting events, as would be typical of natural rain. Biosolids that had been dried to 13% water content (dry-weight basis) increased to 33% water content at the end of a 30-min rewetting period (C.A. Moffet and C.M. Rostagno, unpublished data). At a 90 dry Mg ha–1 rate this equates to 0.18 cm of intercepted water.
Final infiltration rates were markedly increased with the application of biosolids. Sort and Alcaniz (1999) have shown that both fine and coarse soil microporosity were increased with the incorporation of biosolids into soil. This effect on porosity, however, has not been demonstrated for surface-applied biosolids. The increase in infiltration rates that resulted from the application of biosolids in this study was in large part due to an increase in hydraulic conductivity. Soil surface crusts remained throughout the study with cracks, but apparently did not seal (i.e., crust cracks did not close) during rainfall events. The protection due to the biosolids and lower clay dispersibility kept soil particles from detaching and moving into and plugging the cracks. It was occasionally noted that small granules of biosolids accumulated in the cracks between crust polygons. Surface crusts that can form as a result of the slaking action of beating raindrops act as hydraulic barriers impeding infiltration and reducing both initial and steady infiltrability (Hillel, 1982). Surface mulch can serve to intercept and break the impact of the raindrops and thus help to prevent surface sealing (Hillel, 1982).
Rostagno and Sosebee (2001b) cited the "thick layer of ponded water present in the plots treated with biosolids," as supportive of the surface roughness mechanism. The effect of roughness, however, is to decrease the velocity of the runoff water. In the short term, runoff rates from rough plots will be less than from smooth plots while flow depth in the rough plots increases (on plot storage increases). If we assume no difference in infiltration rates between rough and smooth plots, rough plot runoff rates are expected to approach smooth plot runoff rates once on-plot storage due to the roughness is satisfied. Following this logic it is likely that the effect of increased surface roughness on runoff will be greatest during the early phases of a rainfall event and that the differences in runoff rates will diminish with time. Hillel (1982)(p. 224) writes that "the effects of ponding depth and initial wetness can be significant during early stages of infiltration, but decrease in time and eventually tend to vanish in a very deeply wetted profile." The data presented in this study indicated significant differences in runoff extending to the end of a 30-min event when steadystate infiltration is reached, which suggested that the differences in runoff rates are the result of differences in hydraulic conductivity.
The hydraulic head may be slightly increased by hydraulic roughness because velocity of runoff flow slows; however, it is unlikely that the roughness due to the biosolids would cause an increase in the hydraulic head large enough to explain the difference in steady-state infiltration rate. Simultaneous solution of the Green–Ampt model holding wetting front suction and hydraulic conductivity constant between the 0 and 90 dry Mg ha–1 rates in Chilicotal soil required a hydraulic head of 40 cm in the 90 dry Mg ha–1 rate to yield the measured steady-state infiltration rate in both treatments—an unreasonable depth of water. Further, the low hydraulic conductivity of the surface crust in well-crusted rangelands can dominate the infiltration process. Hillel (1982)(p. 227) writes that "failure to account for the formation of a crust can result in gross overestimation of infiltration." Duley (1939) showed that an unprotected surface can have significant crust formation that reduces infiltration. A more likely explanation of the measured difference in steady-state infiltration rate between the biosolids application rate treatments is that the protection afforded the biosolids both physically and due to increased binding of soil particles has altered the hydraulic conductivity of the crust by reducing the potential for the cracks between crust polygons to close and seal.
The runoff response of the Chilicotal soil provides further evidence of increased hydraulic conductivity due to biosolids application. The Chilicotal soil had significant amounts of roughness from 48% surface gravel cover (versus 5% on the bare Stellar soil). The application of biosolids to the Chilicotal soil would impact effective roughness less than application to the Stellar soil because biosolids resting on a pebble would not necessarily confer additional roughness (i.e., the base of the biosolids would not be submerged in the runoff), yet the effect of biosolids application on infiltration was greater in the Chilicotal soil than in the Stellar.
Erosion Response to Biosolids Application
Soil erosion was reduced by the application of biosolids (Table 9). The effect of biosolids application on reducing erosion was greatest where the biosolids were applied to more erodible bare Stellar soils (Fig. 4)
. The Chilicotal soil was devoid of herbaceous vegetation, but the presence of lag gravel armored the soil against erosion. The Stellar soil had both vegetated and bare components. The bare Stellar soil had little gravel on the surface to protect against erosion so even the addition of 7 dry Mg ha–1 of biosolids resulted in a significant reduction in erosion. For each of the three soil–cover combinations studied there were no differences in the erosion measured from plots receiving 34 or 90 dry Mg ha–1 of biosolids (Fig. 4).
|
|
Kladivko and Nelson (1979) and Glauser et al. (1988) suggested that water repellency protects soil units from the stress of wetting and drying. Guidi et al. (1983) reported that soil stability was increased more with fats, waxes, oils, and resins (mostly hydrophobic compounds) extracted from biosolids than with water-soluble polysaccharides. Both soil binding effects and cover effects of biosolids contribute to soil erosion reduction with the biosolids application. Rostagno and Sosebee (2001b) reported that biosolids application reduced clay dispersibility in the upper 3 cm of soil, which would tend to reduce soil erodibility.
|
CONCLUSIONS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
In general, infiltration and steady-state infiltration rates increased and erosion decreased with increasing biosolids application rate, but there is little effect of post-application age on infiltration or erosion over the range of post-application age studied. Winter-season applications of biosolids were hydrologically more effective than summer-season applications. Winter-season biosolids applications had greater water content, but lower Zn, Cu, and Mn content. The connection between biosolids characteristics and hydrological response is unclear. As expected, infiltration and infiltration rates were greater into vegetated than into bare Stellar soil.
|
ACKNOWLEDGMENTS |
|---|
|
NOTES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
REFERENCES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  P. Jurado-Guerra, D. B. Wester, and E. B. Fish Soil nitrate nitrogen dynamics after biosolids application in a tobosagrass desert grassland. J. Environ. Qual., March 1, 2006; 35(2): 641 - 650. [Abstract] [Full Text] [PDF]
|
|
