CONTROLLING SOIL EROSION AND RUNOFF WITH POLYACRYLAMIDE AND PHOSPHOGYPSUM ON SUBTROPICAL SOIL

Transactions of the ASAE
Vol. 48(1): 149−154 2005 American Society of Agricultural Engineers ISSN 0001−2351 149
CONTROLLING SOIL EROSION AND RUNOFF WITH
POLYACRYLAMIDE AND PHOSPHOGYPSUM ON SUBTROPICAL SOIL
B. H. W. Cochrane, J. M. Reichert, F. L. F. Eltz, L. D. Norton
ABSTRACT. Sandy soil, prone to intense soil erosion, is used for agriculture in the subtropics of Brazil. This study was con- ducted to determine whether soil amendments are effective for conserving topsoil by preventing water−induced erosion on a Brazilian sandy Alfisol soil (coarse−loamy, mixed, thermic Typic Paleudalf). A programmable rainfall simulator was used at the experimental station of the Federal University of Santa Maria, in a newly harvested black oat (Avena estrigosa L.) field that was moldboard plowed and disked twice. Plots were on bare tilled soil with 8% to 12% slopes. The soil treatments consisted of a single 5 Mg ha−1 surface application of byproduct phosphogypsum (PG), a single 20 kg ha−1 surface application of anionic polyacrylamide (PAM), a combined amendment (PAM+PG) with the same rates as above, and an unamended soil (control). Simulated rainfall average intensity was 25 mm h−1 with a 2 h duration. Sediment and runoff samples were collected at intervals during the experiment, and soil surface samples inside the plot were taken after the rain for surface crusting analysis. Total soil loss was significantly lower for the treatments than for the control and averaged 197, 278, 217, and 2181 kg ha−1, respectively for PG, PAM, PAM+PG, and control treatments. PAM and PAM+PG had steady−state runoff rates significantly less than that of the control. All of the amendments reduced soil sediment erosion (average 90% reduction) more than final runoff (average 35% reduction). Using amendments to reduce precipitation−induced erosion is a possible alternative conservation practice in this region of the world.
Keywords. Micromorphology, Surface sealing and crusting, Tropical soils, Water erosion, Water infiltration.
he southern states of Brazil have a favorable cli- mate and soil environment for growing crops. The subtropical climate allows ample time for farmers to seed and harvest crops due to the small seasonal
temperature variations commonly found in these regions. In- tensive agricultural production and high−intensity rainfall may cause intense erosion, which is a common limitation for crop production in these soils.
One alternative to reduce erosion is the surface applica- tion of soil amendments, including gypsiferous materials and anionic polyacrylamide. Mined gypsum, gypsiferous materi- als, and phosphogypsum (PG), a byproduct of the phosphate fertilizer industry, decrease erosion by releasing electrolytes to low−electrolyte rainwater and creating a flocculated soil surface condition (Norton, 1995). The effectiveness of PAM and gypsiferous byproducts (PG and others) in controlling erosion and runoff has been shown for several soil types and surface conditions and for simulated and natural rainfall (Miller and Scifres, 1988; Ben−Hur et al., 1992; Reichert and Norton, 1994, 1996; Norton, 1995, Zhang et al., 1998;
Article was submitted for review in December 2003; approved for publication by the Soil & Water Division of ASAE in November 2004.
The use of a trade name, proprietary product, or specific equipment does not imply endorsement by USDA−ARS or Purdue University.
The authors are Belle H. Wallace Cochrane, Research Associate, USDA−ARS National Soil Erosion Research Laboratory and Purdue University, West Lafayette, Indiana; José Miguel Reichert, Professor, and Flávio Luiz Folleto Eltz, Professor, Soils Department, Federal University of Santa Maria, Santa Maria, RS, Brazil; and L. Darrell Norton, Director, USDA−ARS−National Soil Erosion Research Laboratory, West Lafayette, Indiana. Corresponding author: Belle H. Wallace Cochrane, Casilla 6329, Santa Cruz, Bolivia; phone: 591−3−343−5717; e−mail: wallace@agteca. com.
Sirjacobs et al., 2000; Flanagan et al., 2002a, 2002b; Peterson et al., 2002; Yu et al., 2003). However, their effectiveness is dependent on the properties of the amendment and of the soil, as demonstrated for gypsiferous materials by Reichert and Norton (1996). None of the previous studies were conducted in−situ on subtropical, highly weathered, sandy soil.
Both mined gypsum and PG applied to the soil surface increase the electrolyte concentration of rainwater, thus flocculating the clay particles to reduce sediment dispersion and sealing (Norton, 1995). Miller and Scifres (1988) researched the highly weathered Ultisols of the southeastern U.S. and found that gypsum reduced clay dispersion. Smith et al. (1990) measured soil erosion and runoff reductions on a non−calcareous sandy loam soil with application of PG by increasing the electrolyte concentration of the surface solution. Fluidized bed combustion bottom ash improved the infiltration rate and reduced erosion on swelling soils such as Vertisols (Reichert and Norton, 1994) but had a variable effect for highly weathered Ultisols and Oxisols (Reichert and Norton, 1996).
Surface sealing affects the final infiltration rate of a soil. Structural crusting results from raindrop impact crushing the surface aggregates and is described as having a “washed−in” layer and a “washed−out” layer (West et al., 1992; Reichert and Norton, 1995). The washed−out layer consists of larger−grained particles, while the washed−in layer, found underneath, is composed of finer materials that plug the soil pores. Additional- ly, aggregate destruction and surface compaction is enhanced by physio−chemical dispersion. Infiltration rate is significantly affected by the electrolyte concentration and exchangeable sodium percentage of the soil solution near the soil surface (Levy, 1996; Gal et al., 1984).
T
150 TRANSACTIONS OF THE ASAE
Polyacrylamide (PAM) is a synthetic organic polymer that has a long molecule of identical atom chains held together by covalent bonds (Seybold, 1994) that form bridges to soil particles through cations or anions in soil solution. Adsorp- tion of PAM to aggregates is required for this conditioner to be effective. Polymer properties such as charge density, molecular weight, and electrical charge have an effect on soil colloidal reactions (Shainberg, 1992). Repulsion or attrac- tion can occur between the polymer and soil clay particles depending on the charge of each and can result in dispersion or flocculation, respectively. Upon comparing different PAM configurations, Green et al. (2000) found that anionic, 30% charge density, and 12 Mg mol−1 molecular weight character- istics were overall best for different soil types to improve final infiltration rates. Polymer adsorption is also dependent on water quality. Particularly, anionic polymers were found to be more readily adsorbed using a higher electrolyte concentration well water (electrical conductivity (EC) = 0.7 dS m−1) compared to surface water (EC = 0.05 dS m−1) (Shainberg, 1992).
Soil characteristics play a role in adsorption of PAM. Malik and Letey (1991) first indicated that polymers adsorb primarily to the clay sites and exterior of aggregates. Organic matter and clay particles are primary sources of soil electrical charges; therefore, binding was originally thought to occur mostly at these sites. However, recent findings from Levy and Miller (1999) suggest that aggregate adsorption of PAM also occurs with coarse soils. PAM’s large molecular weight makes it impossible to penetrate fine clay aggregates, but coarse soil’s large interparticle space make interior adsorp- tion possible. Intra−aggregate adsorption of PAM and its binding effect suggest that PAM may prevent aggregate destruction from raindrops.
Surface−applied PAM is more effective in reducing rill erosion than runoff in high ESP (exchangeable sodium percentage) soils; in low ESP soils, runoff is more affected than erosion (Levy et al., 1995). Anionic PAM significantly reduced sediment yield on steep slopes of a silt loam soil (Flanagan et al., 2002a, 2002b) and was effective in flocculating clay and stabilizing soils at the surface with high electrolyte concentration water (Ben−Hur et al., 1990). Dontsova (1998) found that PAM decreased erosion of several smectitic soils.
PAM may be useful as an alternative conservation practice to reduce water−induced erosion, thereby conserving top soil in areas where environmental conditions such as high rainfall intensity exists. PAM has been estimated to cost $15 to $20 per kg ha−1, which may be outweighed by savings associated with erosion−related expenses (field operations) and crop responses (higher yields), as reported by Sojka and Lentz (1997) and McCutchan et al. (1993). Ultimately, management choices depend on the overall costs and control effectiveness.
The objective of this study was to determine the effects of PG and PAM on a highly weathered Brazilian Alfisol, in a field setting, on runoff and erosion. It was hypothesized that PAM and PG would benefit these soils by reducing surface sealing/crusting, runoff, and erosion.
MATERIALS AND METHODS Santa Maria is in the Central Depression in the state of Rio
Grande do Sul, in southern Brazil. The landscape is rounded slopes of 8% to 10%, and naturally occurring vegetation is a thick shrubbery. This region has subtropical temperatures, with rainfall fairly well distributed during the year. The soil is a coarse−loamy, mixed, thermic Typic Paleudalf (Soil Survey, 1994). Rainfall erosivity is greatest in October, November, and February (Ministério da Agricultura, 1973). The site was located at the agricultural experimental station of UFSM (Federal University of Santa Maria), and the study was conducted in December 1999. This soil was chosen because of its high rill and interrill erodibility (Reichert et al., 2001).
Soil chemical and physical property analysis of the soil surface (top 5 cm) included particle size analysis using the pipette method (Gee and Bauder, 1986), organic matter percentage using the Walkley−Black method (Nelson and Sommers, 1996), phosphorus content using the colorimetric method (Kuo, 1996), potassium, magnesium, and calcium measurement using neutral ammonium acetate extraction (Helmke and Sparks, 1996), and soil pH in a 1:1 soil:water mixture and 0.01M CaCl2 (Thomas, 1996). X−ray diffraction was performed for identification of the clay fraction to determine clay chemical and physical properties. Phospho- gypsum was analyzed for chemical composition, pH, and X−ray diffraction. The National Soil Erosion Research Lab and Purdue University conducted all of these analyses.
The study was conducted on a newly harvested black oat (Avena estrigosa L.) field with straw removed, following moldboard plowing and disking twice. The field had a history of plowing. Immediately after the second disking, the plots were laid out on the bare soil. Plot size was 1 × 3 m with a slope range of 9% to 12% and an average plot slope of 10%. The following four treatments were randomly assigned to each plot: 5 Mg ha−1 PG, 20 kg ha−1 PAM (anionic, 30% charge density, 12 Mg mol−1), PAM+PG (at the same rates as above), and a control without any amendment. There were three replications for the treatments. PAM was mixed with well water (EC = 0.109 dS m−1) at a concentration of 200 ppm and applied 24 h before the simulated event to allow it to dry. However, several of the PAM−applied plots did not appear completely dry upon initiation of rainfall simulation due to the amendment application, which required 10 mm of water for a useable viscosity for surface application. A moist appearance was inconsistent within and among the PAM−ap- plied plots. The field plots were covered at night to prevent possible infiltration of natural rainfall; however, natural field moisture variations may have impeded the drying process of some PAM plots. The soil surface moisture of all the plots averaged 4% to 9% in the 0 to 10 cm surface section. The PG was applied immediately prior to the trial using a sieve. Neither the PG nor the control plots were prewetted before the simulated rainfall.
A Norton−type programmable rainfall simulator (unpub- lished) was used in the field and centered over two plots with angled plot end barriers catching the soil and runoff water. The rainfall rate was measured by rain gauges placed around the plots, and averaged 25 mm h−1 with a ±5% variation. Well water was used in the rainfall simulator (EC = 0.11 dS m−1; pH = 6.15 at 25°C). After installation of the barriers, the rainfall simulator was centered, and trenches were dug to
151Vol. 48(1): 149−154
prevent spillover from the upslope plot into the lower one. Well water rainfall was applied for 2 h. The time to runoff initiation was noted, and runoff samples were collected. After the 2 h rain, plot slope was measured. After a 24 h period, a sample of the soil surface was taken for micro- morphological analysis.
The runoff samples were collected from the plot end in 5 min intervals, measuring the time required to fill a 1 L bottle. Afterward, the samples were weighed, alum (AlK(SO4)2·12H20) was added to flocculate suspended soil, the water portion was decanted, and the remaining soil was oven dried at 100°C and then re−weighed to determine sediment concentration and average soil and water loss values. Undisturbed soil surface samples were collected for image analysis. The samples were allowed to air dry for two weeks, placed in an oven for several days, and impregnated with an epoxy resin.
All treatments were replicated three times in a complete randomized experimental design. A regression of runoff measurements was performed in order to fit the field data, using curve−fitting software (Jandel Scientific, 1992). The data were analyzed using the SAS ANOVA (SAS, 1998) procedure and least significant difference test at the = 0.05 level and Tukey’s test at p < 0.05.
RESULTS AND DISCUSSION SOIL, WATER, AND PHOSPHOGYPSUM PROPERTIES
On the soil surface (top 5 cm), 64% of the soil was sand−size particles, with 20% of that in the medium (0.25 to 0.15 mm) size class, and clay content was 8%. The upper horizon of this soil was formed from the alluvial deposition of parent materials from the higher slopes and is composed of sandstone with siltstone and shale at greater depths (Ministério da Agricultura, 1973).
The soil cation exchange capacity of 7.7 cmol kg−1 in the top horizon is typical of soils with high sand content and low soil organic matter content (1.8%). Soil pH in water was 5.4, and pH in 0.1M CaCl2 solution was 4.7, suggesting a net negative charge. The P content was 18 mg kg−1, K was 73 mg kg−1, Mg was 162 mg kg−1, and Ca was 825 mg kg−1. The clay fraction mineralogical analysis using X−ray diffraction indicated a kaolinite−illite mixture in the 0 to 5 cm soil layer with a small amount of variable charge gibbsite.
The electrolyte concentration of the well water used for the simulated rainfall and to mix the PAM for application was 0.109 dS m−1. It is likely that this water was an electrolyte source that aided in PAM adsorption to the coarse soil.
The PG used was a byproduct of a phosphorus fertilizer plant in the state of Rio Grande do Sul, Brazil. Samples were tested and compared to Resource Conservation and Recovery Act waste application permit limits, showing large quantities of calcium, sulfur, iron, and phosphorus (table 1). X−ray diffraction of the PG showed that the mineralogy of the material is largely composed of mined gypsum (CaSO4·2H20), and apatite (Ca2Fe(PO4)2·2H20) formed dur- ing the byproduct fabrication process. The pH of PG in water was 4.8, meaning the dissolution reaction creates an acidic environment from the breakdown of the various minerals.
Table 1. Phosphogypsum chemical properties compared to Resource Conservation and Recovery Act permit limits.
Element Type Chemical Sample
(mg kg−1)
RCRA Permit Limits
(mg kg−1)
RCRA metals Arsenic n.d. 41 Cadmium n.d. 39 Chromium 1.3 NL Lead 8.8 300 Nickel 3.3 420 Selenium n.d. 100 Mercury n.d. 17 Copper 56.0 1500
Other elements Molybdenum 1.2 NL Zinc 4.8 2800 Phosphorus 370.0 NL Ammonium−N n.d. NL Sodium 80.0 NL Magnesium 60.0 NL Potassium n.d. NL Sulfur 18500.0 NL Calcium 23010.0 NL Iron 1152.0 NL Aluminum 3.0 NL Manganese 4.0 NL Cobalt 5.0 NL
NL = no limit. n.d. = not detected.
RUNOFF AND INFILTRATION RATE
A typical runoff curve was observed for most treatments, with initially no runoff, then an increase in runoff (otherwise known as the transition period), and finally reaching field equilibrium (known as steady−state runoff). Figure 1 depicts runoff rates (mm h−1) over time from the start of rainfall. The measured averages per treatment are indicated by symbols, with the solid lines representing regression fit of the data. From the runoff hydrograph, several key points of discussion can be identified: the amendments’ abilities to lower final runoff rates, the effect of the amendments on cumulative runoff, the unusual runoff curve of PAM+PG, and finally the differing times to runoff.
For all the treatments, the final runoff rates indicated on the graph show the amendments’ abilities to lower field runoff. The runoff curves were fit through a regression analysis. After 2 h of simulated rain, the control plots ended with higher runoff rates than any other treatment. Final runoff and infiltration rates (table 2) were calculated by averaging the last three measurements and represent steady−state conditions for all the treatments except for the PAM+PG treatment, which had not yet reached steady state after the timed 2 h rainfall. Final runoff rates were significantly lower for the PAM and PAM+PG treatments than for the control (Tukey’s p < 0.05), while PG did not produce statistically significant results. Final runoff values were reduced by 45%, 36%, and 31% compared to the control for the PAM, PAM+PG, and PG treatments, respectively. The control shows characteristically higher runoff rates and lower infiltration rates than the treated soils, which is confirmed by phosphogypsum research on Ultisols conducted by Miller (1987) and PAM research by Flanagan et al. (1997).
Cumulative runoff (table 3) does not show any significant differences among the treatments using the LSD test at the
152 TRANSACTIONS OF THE ASAE
0
2
4
6
8
10
12
14
16
18
20
time (min)
ru n
o ff
(m m
/h )
PAM +PG PG PAM control PAM + PG fit PG fit PAM fit Control fit
0
200
400
600
800
1000
1200
1400
1600
time (min)
se d
im en
PAM + PG PG PAM Control
Figure 1. Runoff rate (upper) and cumulative soil loss (lower) vs. time from initial rainfall for control and surface amendments. For runoff rates, sym- bols are averaged measured data, and lines are regressions of measured data. PG is phosphogypsum and PAM is polyacrylamide.
Table 2. Time to runoff, final runoff, and final infiltration.[a]
Treatment[b]
Final Runoff Rate
(mm h−1)
Final Infiltration Rate
(mm h−1)
Control 53 ab 14.6 b 10.7 b PAM 44 b 8.1 a 17.6 a PG 93 a 10.1 ab 15.6 ab
PAM+PG 8 c 9.3 a 16.4 a [a] Means followed by the letter, in a given column, do not differ statistical-
ly (Tukey’s test, p < 0.05). [b] PG is phosphogypsum; PAM is polyacrylamide.
Table 3. Total runoff, total soil loss, and average sediment concentration.[a]
Treatment[b]
Average Sediment Conc.
(g L−1)
Control 10.2 ab 2180.8 a 21.4 a PAM 5.1 b 277.8 b 5.5 b PG 3.6 b 197.4 b 5.5 b
PAM+PG 18.1 a 217.4 b 1.2 b [a] Means followed by the letter, in a given column, do not differ statistical-
ly (LSD 5%). [b] PG is phosphogypsum; PAM is polyacrylamide.
153Vol. 48(1): 149−154
5% level. Total runoff values were 3.6, 5.1, 18.1, and 10.2 mm, respectively, for the PG, PAM, PAM+PG, and control. However, figure 1 shows that the PG and PAM treatments lowered the total runoff compared to the control, while PAM+PG may have induced more runoff initially.
PAM+PG cumulative runoff results from this research do not coincide with other studies on the combined amend- ments. Yu et al. (2003) treated a sandy clay Alfisol with dry granule PAM and gypsum and found increased infiltration rates. As shown in figure 1, the PAM+PG treatment did not display typical runoff characteristics: the runoff rate was relatively the same throughout the entire rainfall period. This resulted in the highest cumulative runoff of all the treatments. It was observed during the field experiments that a slick surface developed on the PAM+PG treated soil, which may have contributed to the constant high runoff. Further research is needed to quantify the interactions of the PAM+PG treatment as related to the…