Agronomy Journal Journal of Natural Resources and Life Sciences Education
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published online 5 September 2006
Published in Agron J 98:1298-1311 (2006)
DOI: 10.2134/agronj2005.0184
© 2006 American Society of Agronomy
677 S. Segoe Rd., Madison, WI 53711 USA
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Izaurralde, R. C.
Right arrow Articles by Quiroga Jakas, M. C.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Izaurralde, R. C.
Right arrow Articles by Quiroga Jakas, M. C.
Agricola
Right arrow Articles by Izaurralde, R. C.
Right arrow Articles by Quiroga Jakas, M. C.
Related Collections
Right arrow Sustainable Agriculture
Right arrow Best Management Practices
Right arrow Nutrient Management
Right arrow Soil Conservation
Right arrow Other Soil Management

Production Papers

Crop Performance and Soil Properties in Two Artificially Eroded Soils in North-Central Alberta

R. C. Izaurraldea, S. S. Malhib,*, M. Nyborga, E. D. Solbergc and M. C. Quiroga Jakasa

a Dep. of Renewable Resources, 4-42 ESB, Univ. of Alberta, Edmonton, AB T6G 2E3, Canada (R.C. Izaurralde, present address: Pacific Northwest National Lab., 8400 Baltimore Ave., Suite 201, College Park, MD 20740)
b Agriculture and Agri-Food Canada, Melfort, SK S0E 1A0, Canada
c Alberta Agriculture, Food & Rural Development, Crop Diversification Centre North, 17507 Fort Road NW, Edmonton, AB T5Y 6H3, Canada

* Corresponding author (malhis{at}agr.gc.ca)

Received for publication June 20, 2005.

   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Field experiments were conducted from 1991 to 1995 at Josephburg (Orthic Black Chernozem, Typic Cryoboroll) and Cooking Lake (Orthic Gray Luvisol, Typic Cryoboralf), Alberta, to determine the impact of topsoil removal on selected soil properties, N-mineralization potential, and crop yield, and the effectiveness of various amendments for restoring the productivity of eroded soils. The simulated-erosion levels were established in the autumn of 1990 by removing 20 cm of topsoil in 5-cm depth increments. The four amendments were: control, addition of 5 cm of topsoil, fertilizers to supply 100 kg N ha–1 and 20 kg P ha–1, and cattle manure at 75 Mg ha–1. Topsoil and manure were applied once in the autumn of 1990, while fertilizers were applied annually from 1991 to 1995. Available N and P; total C, N, and P; and N-mineralization potential decreased, while bulk density increased with increasing depth of topsoil removal. Tiller number, plant height, spike density, thousand-kernel weight, and leaf area index decreased with simulated erosion. Grain yield reductions due to simulated soil erosion were either linear or curvilinear functions of nutrient removal. Application of N and P fertilizers and manure improved grain yield and reduced the impact of yield loss due to erosion. Return of 5 cm of topsoil also increased grain yield, but to a lesser extent than manure or fertilizers. Grain yields were maximized when fertilizers were also applied to organic amendment treatments. Our findings suggest the importance of integrated use of organic amendments and chemical fertilizers for best crop yields on severely eroded soils.

Abbreviations: ASW, available soil water • Db, bulk density • LAI, leaf area index


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 
TOPSOIL thickness is a major indicator of soil quality and productivity (Power et al., 1981; Christensen and McElyea, 1988; Larney et al., 2000, 2003). The physical, chemical, and biological properties of this surface horizon govern the reception, storage, and transfer of water and energy. Topsoil stores most of the available nutrients present in a soil profile. Erosion and deposition processes caused by wind and water as well as tillage implements may affect the thickness of topsoil and hence its productivity.

In western Canada, soil erosion remains one of the most important processes potentially affecting the productivity of agricultural lands (Coote, 1984; Sparrow, 1984). Eroded soils decrease plant yield through increased bulk density, poorer tilth, and reduced organic matter content, nutrient availability, and water-holding capacity (Batchelder and Jones, 1972; Power et al., 1981; Frye et al., 1982; Dormaar et al., 1986; Tanaka and Aase, 1989; Arriaga and Lowery, 2003). The altered properties of eroded soils reduce plant growth by altering root density patterns, crop growth rates, and developmental stages. Reduced crop growth and development create, in turn, opportunities for greater weed growth.

The study of crop productivity as a function of erosion is complex because soil and topographic factors interactively influence crop yield (Langdale and Schrader, 1982; Langdale et al., 1985; Whitman et al., 1985; Daniels et al., 1985; Schertz et al., 1985a, 1985b; den Biggelaar et al., 2001, 2003). The effects of erosion on crop productivity are hardly detectable during the early stages of the erosion processes because yield reductions are usually small (Bakker et al., 2004; Govers et al., 2004). When the signs of soil erosion (e.g., dunes, rills, and gullies) become evident, however, it is already too late to use preventive measures alone (Lyles, 1975; Battiston et al., 1985; Frye et al., 1982). Thus, the application of corrective measures (e.g., terraces) to reduce erosion or repair damage becomes more costly and labor intensive. Yield differences between eroded and uneroded soils, however, are often masked by crop genotypes that respond well to application of fertilizers and herbicides (Krauss and Allmaras, 1982).

Nitrogen and P fertilizers have been used to restore the productivity of artificially exposed subsoil (Carlson et al., 1961; Engelstad and Shrader, 1961; Morrison and Shaykewich, 1987; Shafiq et al., 1988; Tanaka and Aase, 1989) but their effectiveness varies with soil type, climate, crop, and level of management (Eck, 1968, 1969; Langdale and Schrader, 1982; Morrison and Shaykewich, 1987; Arce-Diaz et al., 1993). Research in greenhouse and field experiments has shown that even though crop yields under eroded-soil conditions can be increased with the addition of N and P fertilizers, the levels obtained under slightly or uneroded conditions may not be realized (Frye et al., 1982; Massee and Waggoner, 1985; Mielke and Schepers, 1986; Eck, 1987; Malhi et al., 1994; Larney et al., 1995b; Izaurralde et al., 1998b) due to deterioration of tilth and physical, chemical, and biological properties of the soil.

Both wind and water erosion are important processes occurring on Alberta landscapes (Tajek et al., 1985). Relatively few studies have been conducted to assess the effects of erosion on the productivity of agricultural lands in Alberta. Dormaar et al. (1986) used simulated-erosion plots in southern Alberta to assess methods to restore the productivity of eroded soil. Although their findings are important, these studies did not provide a direct method for quantifying the effects of erosion on crop productivity as it would occur under natural environments (Bakker et al., 2004; Govers et al., 2004). Quantification of erosion–productivity relationships would provide improved cost–benefit analysis of conservation tillage and erosion control programs.

Three approaches have been used to quantify the effects of erosion on soil productivity (Meyer et al., 1985; Lal, 1988): (i) the assessment of past erosion effects on soil productivity (Schertz et al., 1985b; McDaniel and Hajek, 1985; Daniels et al., 1987); (ii) the influence of simulated erosion on crop yield (Mbagwu et al., 1984; Morrison and Shaykewich, 1987); and (iii) simulation of the long-term effect of erosion on crop productivity (Williams et al., 1983). We used a simulated-erosion approach (i.e., artificial removal of topsoil; Dormaar et al., 1986; Tanaka and Aase, 1989; Ives and Shaykewich, 1987; Morrison and Shaykewich, 1987; Yost et al., 1985) to determine the short-term impact of topsoil removal on selected soil properties, N-mineralization potential, and crop yield, and the effectiveness of various amendments for restoring the productivity of the eroded soils.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Field Experiments
Field experiments were conducted from 1991 to 1995 at Josephburg and Cooking Lake, both located ~30 km east and southeast, respectively, of Edmonton, AB (53°34' N, 113°33' W) and separated by the same distance from each other. Meteorological information for Josephburg was obtained from the Edmonton Namao Airport station while that for Cooking Lake was derived from the Edmonton International Airport station. The soil at Josephburg is an Orthic Black Chernozem (Typic Cryoboroll) of the Angus Ridge series on a 2% slope with 30-cm-thick A horizon (Ap + Ah). An Orthic Gray Luvisol (Typic Cryoboralf) of the Cooking Lake series on a 2% slope is found at the second site. The thickness of the Ap horizon of the Orthic Gray Luvisol is only 15 cm. Both soils lie on glacial-till parent material and have been under cultivation for >40 yr. The common crop rotation practiced at Josephburg before the initiation of the experiment was canola (Brassica spp.), wheat (Triticum aestivum L.), and barley (Hordeum vulgare L.). In recent years, N fertilizer was usually applied in the autumn at a rate of 60 kg N ha–1 while P fertilizer was applied with the seed at a rate of 15 kg P ha–1. In contrast, at Cooking Lake oat (Oryza sativa L.) was intercropped with field pea (Pisum sativum L.) in rotation with barley. Cattle manure at ~5 Mg ha–1 yr–1 was applied to this field. In addition, the field was annually fertilized with N at 60 kg ha–1 and P at 10 kg ha–1.

The experimental design was a split plot, replicated four times, with five erosion levels randomly arranged as main plot and four amendments assigned as subplot treatments. The simulated-erosion levels (0, 5, 10, 15, and 20 cm) or cuts were established in the autumn of 1990 by removing up to 20 cm of topsoil using an excavator with a grading bucket in 5-cm depth increments. The four amendments were: (i) control (C), (ii) addition of 5 cm of topsoil (T), (iii) addition of fertilizer (F) to supply 100 kg N ha–1 (as urea sidebanded) and 20 kg P ha–1 (as triple superphosphate applied with seed), and (iv) cattle manure (M) at 75 Mg ha–1 on a dry-weight basis. The topsoil and manure amendments were applied only once soon after the erosion levels were established in the autumn of 1990, while the fertilizer amendment was added in the spring annually from 1991 to 1995.

The main plots measured 10 by 14.6 m while the subplots were 10 by 3.65 m. In 1992, the main experiment was modified to further test the residual effects of manure and topsoil amendments. Each experimental unit for the manure and topsoil amendments was split in two. One half did not receive any fertilizer; the other half received 100 kg N ha–1 as broadcast-incorporated urea and 20 kg P ha–1 as triple superphosphate applied with the seed. At each site, the resulting experimental design was a split-split plot with four replications. All plots were sown to hard red spring wheat (cv. Roblin) at a seeding rate of 90 kg ha–1 in early May every year. Weeds were controlled with herbicides and by hand.

Plant Growth and Yield Measurements
Plant variables measured included: (i) leaf area index (LAI) with a LI-COR 2000 Plant Analyzer (LI-COR, Lincoln, NE), (ii) plant density, (iii) growth stages using the Zadock's scale (Zadocks et al., 1974), (iv) height at maturity, (v) tiller and spike density, (vi) aboveground biomass at anthesis, (vii) grain and total dry matter (seed + straw), yield, and (ix) thousand-kernel weight.

Plant population was measured 4 wk after seeding, plant height was measured at maturity, and tiller density was calculated by dividing spike density by plant density. The plots were harvested with a plot combine (11.3 m2, Wintersteiger, Salt Lake City, UT) in late August or early September. Representative grain subsamples were collected to determine moisture to adjust grain yields to 12% moisture content. Total nutrient mass was calculated using mass values of grain and straw determined after drying at 60°C. Samples of grain and straw were then ground to pass a 1-mm sieve and prepared for nutrient analysis. Nitrogen and P concentrations were determined on an acid digest using a Technicon autoanalyzer (Technicon Industrial Systems, 1977a).

Soil samples for root mass and length were obtained in late August 1991 following procedures detailed in Izaurralde et al. (1992, 1993). Briefly, soil cores for root mass and length determinations were obtained with a root auger (10-cm diam., Eijkelkamp J-199A, Amsterdam, the Netherlands) in August 1991 from the uneroded and 20-cm eroded plots in the control, fertilizer, and manure treatments at both locations. Four subsamples (two in-row and two in midrow positions) were taken in 10-cm increments to a depth of 40 cm. The intact soil cores were kept frozen at –20°C until processing. At this time, the samples were thawed and the roots separated from the soil using a hydropneumatic elutriation method (Smucker et al., 1982). Washed root samples underwent further processing to separate organic debris from roots. Root weights were obtained from samples oven dried at 50°C for 72 h. Root length was determined using the line-intersect method of Tennant (1975) with digitizing followed by microcomputer image analysis.

The data on plant density, tiller number, plant height at maturity, spike density, aboveground biomass at anthesis, grain yield, total dry matter yield, root mass, and thousand-kernel weight were subjected to ANOVA using the GLM procedure of SAS (SAS Institute Inc., 1989). Fisher's least significant difference (LSD0.05) was used to determine significant differences among treatment means. Leaf area index development during the growing season was estimated with PROC NLIN in SAS (SAS Institute Inc., 1989) by fitting a logistic-growth model function [y = a/(1 + be–cx), where y is LAI (m2 m–2); a, b, and c are regression coefficients; and x is days after seeding] to the observed data.

Soil Measurements
Four soil cores (4.2-cm diam.) per plot were taken in early May 1991 from all plots, divided into five depth increments (0–10, 10–20, 20–30, 30–60, and 60–90 cm), and bulked by depth and plot. Soil samples were air dried and ground to pass a 2-mm sieve. Physical analyses of soil samples included texture and water-holding capacity at 33 and 1500 kPa. Available soil water to 90-cm depth (ASW, mm) was calculated with water-holding capacity and bulk density (Db) information. Bulk density was determined in situ for the 0- to 10-cm layer with an MC1 gamma probe (Campbell Pacific Nuclear Corp., Martinez, CA). The Db values for deeper soil layers were obtained from the Soil Inventory Map Attribute File—Alberta, Soil Layer Digital Data (Agriculture Canada, 1989).

Soil chemical analyses included: pH in 1:2 soil/CaCl2 solution, exchangeable bases, extractable and total P, available N (NO3–N), total N, and organic C. Exchangeable Ca, Mg, Na, and K ions were extracted in 1.0 M NH4OAc solution buffered at pH 7 using a 1:12 soil/solution ratio. Cation concentration in the extract was determined by inductively coupled plasma–atomic emission spectrometry. Phosphorus was extracted using a modified procedure of Miller–Axley (Miller and Axley, 1956) and determined colorimetrically on a Technicon Autoanalyzer (Technicon Industrial Systems, 1977a). Nitrate-nitrogen was extracted using a 1:5 soil/2 M KCl solution and its concentration in the extracts was measured with a Technicon Autoanalyzer II (Technicon Industrial Systems, 1977b). Total N and P were determined by Kjeldahl digestion as described by McKeague (1978). Total C was determined using a CR12 Leco Carbon Determinator and was taken as total organic C due to absence of carbonates in the soil samples.

Nitrogen Mineralization Soil Incubation Experiment
Fifteen-centimeter-deep samples were taken in November 1990 from soil profiles artificially eroded in increments of 0, 5, 10, 15, and 20 cm. The soil samples were dried at room temperature and ground to pass a 2-mm sieve. The incubation samples consisted of 20 g each of soil and acid-washed sand, mixed and placed into plastic leaching containers. The experimental units (incubation vessels) were arranged in a randomized complete block design with four replications. The samples were incubated at 30°C and 55% water-filled porosity and leached with a N-free solution at 7, 14, 28, 42, 56, 84, 112, and 168 d (Stanford and Smith, 1972). Ammonium- and NO3–N were determined on the leachates (Technicon Industrial Systems, 1977b) and the cumulative values used to fit one-pool negative exponential models of the form Nm = No[1 – exp(–kt)], where Nm is the N mineralized during time t and No is the N mineralization potential. The statistical procedure used for model fitting was the Marquardt option of PROC NLIN in SAS (SAS Institute Inc., 1989). The error between predicted (PRED) and observed (OBS) values for n observations was evaluated with the RMSE formula:

Formula 1[1]


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Environmental Conditions
At both locations, May to August precipitation was below normal during all years except 1994, when it was above normal (Table 1). The driest growing season was 1992, which only received about one half the normal amount. At both locations, June and August were the wettest months in 1994. In general, precipitation amounts and distribution at the two airports were rather similar, conditions that we presume also extended to the two experimental sites.


View this table:
[in this window]
[in a new window]
  		Table 1. Monthly and total precipitation during 1991 through 1995 and the 1971 to 2000 long-term average at Edmonton Namao Airport (Josephburg) and Edmonton International Airport (Cooking Lake) in Alberta.

 
Soil Properties of Control Plots after Topsoil Removal
Chemical soil properties changed drastically with depth of layer sampled and depth of simulated erosion for both sites (Table 2). For example, the concentration of total C in the upper 10 cm soil of the 20-cm cut at Josephburg was 60% of that in the normal profile, and at Cooking Lake it was only 25% of that in the normal profile. Available N (NO3–N) and extractable P in the soil decreased with increasing depth of topsoil removal at both sites, and the decline was more pronounced at Cooking Lake than at Josephburg. Total N and P in the soil also decreased with increasing depth of topsoil removal, the effects being more pronounced for total N than total P. The cation exchange capacity of the soil decreased from 33.0 to 19.2 cmol kg–1 in the Josephburg soil, but it usually increased in the Cooking Lake soil with increasing depth of topsoil removal (data not shown). At Josephburg, simulated erosion did not greatly influence ASW to the 90-cm depth (data not shown). Available soil water ranged from 140 to 150 mm and Db in the upper 10 cm of soil increased with the depth of simulated erosion. Apparently, most of the expected reduction in ASW due to topsoil removal was compensated by an increase in Db.


View this table:
[in this window]
[in a new window]
  		Table 2. Some characteristics of soils in the field experiments after applying various levels of simulated erosion at two sites in Alberta.

 
Nitrogen Mineralization Kinetics
For the Josephburg soil, the removal of topsoil up to 10 cm did not cause any reduction in mineralized N, but the subsequent increments of topsoil removal resulted in a large decrease of mineralized N (Table 3). For the Cooking Lake soil, cumulative mineralized N decreased with increasing depth of topsoil removal. The values of N-mineralization potential (No) decreased linearly (y = –5.8x + 189.2; R2 = 0.99 at {alpha} = 0.01) with depth of simulated erosion at Josephburg and quadratically (y = 0.6743x2 – 28.966x + 342.11; R2 = 0.97 at {alpha} = 0.01) at Cooking Lake (Table 3). Rate constants increased and then decreased with depth of simulated erosion. Root mean square errors were lowest for uneroded soil but increased substantially with depth of simulated erosion (Table 3).


View this table:
[in this window]
[in a new window]
  		Table 3. Observed and predicted cumulative N mineralization in two artificially eroded soils in Alberta. Predicted cumulative N mineralization values were estimated with a single negative exponential model of the form Nm = No[1 – exp(–kt)]{dagger}.

 
Yield, Yield Components, and Nutrient Concentration and Uptake in 1991
Grain yields were affected significantly by simulated erosion levels and amendments at both sites (Table 4). Average grain yield on uneroded soil was similar at both sites, but it decreased more sharply with increasing erosion at Cooking Lake than at Josephburg. The interaction effect between level of simulated erosion and amendment was significant at Josephburg but not at Cooking Lake.


View this table:
[in this window]
[in a new window]
  		Table 4. Grain yield response of wheat to topsoil removal (cut) and amendments in 1991 at two sites in Alberta.

 
At both sites, commercial fertilizer was the best amendment to restore the productivity of eroded soil, followed by manure and topsoil (Table 4, Fig. 1 ). At Josephburg, grain yields obtained on fertilized plots subject to 20 cm of simulated erosion (3.69 Mg ha–1) were considerably greater than those obtained on uneroded control plots (2.61 Mg ha–1). At Cooking Lake, however, grain yields obtained on fertilized plots subject to 20 cm of simulated erosion (2.67 Mg ha–1) were only slightly greater than those obtained on uneroded control plots (2.55 Mg ha–1). Grain yield for the manure-amended 20-cm cut treatment was 3.44 Mg ha–1 at Josephburg and 1.77 Mg ha–1 at Cooking Lake. The corresponding value for the topsoil-amended 20-cm cut treatment was 1.58 Mg ha–1 at Josephburg and 1.20 Mg ha–1 at Cooking Lake.


Figure 1
View larger version (12K):
[in this window]
[in a new window]
  		Fig. 1. Immediate effects of topsoil removal (cut) and amendments on grain yield of wheat in 1991 at two sites in Alberta.

 
As for grain yield, simulated erosion and amendments had marked effects on total dry matter yield (TDMY) of wheat at both sites (data not shown). The decline in TDMY with increasing erosion was more pronounced at Cooking Lake than at Josephburg (from 6.72 to 1.21 Mg ha–1 at Cooking Lake and 6.84 to 2.52 Mg ha–1 at Josephburg). The increase in TDMY with the addition of amendments differed with amendment, simulated erosion level, and soil type. At both sites, commercial fertilizer was the best amendment to improve TDMY of eroded soil. At Josephburg, the TDMYs in the 0- to 20-cm cuts ranged from 12.36 to 9.07 Mg ha–1 for chemical fertilizer, from 10.42 to 8.75 Mg ha–1 for manure, and from 8.67 to 3.52 Mg ha–1 for topsoil treatment. The corresponding values at Cooking Lake ranged from 11.63 to 6.87 Mg ha–1 for commercial fertilizers, from 10.75 to 4.76 Mg ha–1 for manure, and from 9.11 to 3.27 Mg ha–1 for topsoil treatment.

At Cooking Lake, plant densities were similar in all treatments (Table 5). At Josephburg, plant density in the manure and fertilizer treatments was lower than in the control, with the topsoil amendment showing intermediate values. Interaction between the depth of simulated erosion and amendment was significant at P ≤ 0.059. At neither site did simulated erosion affect final plant stands.


View this table:
[in this window]
[in a new window]
  		Table 5. Effect of topsoil removal (cut) and amendments on growth parameters of wheat in 1991 at two sites in Alberta.

 
Both depth of cut and amendments caused significant differences in plant height at both locations (Table 5). Plant height decreased with the increase in depth of topsoil removal, and the addition of amendments increased it. On average, wheat plants were 10 cm taller at Josephburg than at Cooking Lake.

Averaged across amendments, simulated erosion reduced the number of tillers (Table 5). The 5-, 10-, 15-, and 20-cm cuts reduced spike density by 7.1, 24.7, 42.0, and 41.5%, respectively, at Cooking Lake and by 3.2, 4.1, 16.5, and 23.2%, respectively, at Josephburg. Apparently due to 7.4% lower plant density than the control treatment, the fertilizer and manure treatments increased tillering and hence spike density. The addition of topsoil increased tillering, but not spike density. At both sites, late tillering in unfertilized plots accounted for <8% of spike density. In contrast, late tillering in fertilized plots under 10, 15, and 20 cm of simulated erosion was two- to threefold greater than in corresponding unfertilized plots. Maximum late tillering occurred at Cooking Lake, i.e., 29 and 26% for the 15- and 20-cm cuts, respectively.

At both sites, the differences in LAI among treatments started to appear by 39 d after seeding (DAS; Fig. 2 ). From 46 DAS, the LAI means averaged across erosion levels differed significantly among amendments. The order of LAI was 0-cm cut >10-cm cut ≥20-cm cut. At Cooking Lake, the manure treatment produced the greatest LAI, followed by the fertilizer, topsoil, and control treatments. At Josephburg, however, the LAIs in the 0-cm cut for the last three dates tended to be lower on manured plots than on fertilized plots.


Figure 2
View larger version (31K):
[in this window]
[in a new window]
  		Fig. 2. Observed and predicted LAI (leaf area index) of wheat in the 1991 growing season with different topsoil removals (cut) and amendments at two sites in Alberta.

 
At both sites, inflorescence emergence and anthesis were not reached evenly, but growth-stage differences among erosion levels were minimal after anthesis (data not shown). Clear differences in growth stages developed at Cooking Lake by 46 DAS. By this date, most plants had between 1/4 and 1/2 of their inflorescence emerged except those growing on 0- and 5-cm-cut manured plots, which only had reached the last stage of booting. By 55 DAS, most plants had completed anthesis except those grown on 0- and 5-cm-cut manured plots, which were between the beginning and hlafway through anthesis. At Josephburg, simulated erosion advanced wheat development. The manure treatment reduced variations in wheat development. Wheat development on 20-cm-cut, manure-amended plots was delayed with respect to the other amendments.

At both sites, seeds obtained from the 0- and 5-cm-cut plots were about 6% heavier than those from the 15- and 20-cm-cut plots (Table 5). At Cooking Lake, the manure effects on seed weight were similar to those of fertilizer and topsoil amendments. Both of these treatments produced seeds that were heavier than those obtained from the control plots. A greater weight variability was observed at Josephburg, where manure produced the heaviest seeds.

At both sites, the kind of amendment used had a more pronounced effect on grain N concentration than did the erosion level (Table 6). At Josephburg, the difference in grain N concentration was 2.8 g kg–1 for level of erosion and 7.3 g kg–1 for amendments. The corresponding values at Cooking Lake were 1.0 and 7.1 g kg–1, respectively. The decrease in grain N concentration with the level of simulated erosion was accompanied by a decrease in total N uptake. The application of commercial fertilizers at Josephburg resulted in the highest grain concentration and total uptake of N. Grain N concentration in manure- and topsoil-treated plots was lower than that in control plots. At Cooking Lake, wheat plants grown on fertilized plots took up the highest amounts of N but had a lower grain-N concentration than either control plots at the same site or fertilized plots at Josephburg.


View this table:
[in this window]
[in a new window]
  		Table 6. Effect of topsoil removal (cut) and amendments on concentration and uptake of total N in grain of wheat in 1991 at two sites in Alberta.

 
At Cooking Lake, in the control plots, root mass was reduced markedly with removal of 20 cm of topsoil (data not shown). Total root mass to a depth of 40 cm in the 20-cm eroded plots was 57% that of uneroded plots (473 vs. 730 kg ha–1). Fertilizer application increased root mass substantially in both uneroded (934 kg ha–1) and eroded plots (913 kg ha–1). The addition of manure also increased root mass in both eroded (782 kg ha–1) and uneroded (748 kg ha–1) treatments, but the increase was smaller than in the fertilizer treatment. The response of root length to topsoil removal and amendments had a similar trend to that of root mass. For example, the total root length to a depth of 40 cm in the 20-cm eroded vs. uneroded plots was 2.65 vs. 1.02 km m–2 in the control, 2.85 vs. 2.57 km m–2 in the fertilized, and 2.49 vs. 1.88 km m–2 in the manured plots. At Josephburg, both root mass and length decreased when topsoil was removed but the effects were much less pronounced than at the Cooking Lake site (data not shown).

Erosion and Amendment Effects on Grain Yield in 1992 to 1995
Grain yields declined significantly with topsoil removal in all 4 yr at Cooking Lake, and in 1992 and 1993 at Josephburg (Table 7, Fig. 3 ). The decline in grain yield due to simulated erosion was much more drastic at Cooking Lake than at Josephburg. Grain yield increased markedly with chemical fertilizers in all years at both sites. At Cooking Lake, application of fertilizers to plots previously amended with manure or topsoil increased grain yield further, and grain yields were greater on manured plots than on topsoil-amended plots. At Josephburg, there was little or no additional improvement in grain yield with combined application of fertilizers and organic amendments compared with fertilizers alone.


View this table:
[in this window]
[in a new window]
  		Table 7. Grain yield response to topsoil removal (cut) and amendments in 1992 to 1995{dagger} at two sites in Alberta.

 

Figure 3
View larger version (31K):
[in this window]
[in a new window]
  		Fig. 3. Short-term effects of topsoil removal (cut) and amendments on grain yield of wheat during 1992 through 1995 at two sites in Alberta.

 
At Cooking Lake, there was a significant residual effect of manure and topsoil (added in the autumn of 1990) on grain yield in all 4 yr and the effect on grain yield was greater with manure than topsoil (Table 7, Fig. 4 ). At Josephburg, the residual effect of manure on grain yield was significant in 3 of the 4 yr, but with topsoil amendment it was significant only in 1993. The interaction effects of cut x organic amendment on grain yield at Cooking Lake indicated that the residual effects of manure and topsoil addition decreased with increasing depth of simulated erosion, with the least effect for the 20-cm cut (Fig. 4). The trend was similar for the Josephburg site, but the differences between simulated erosion levels were smaller than that at Cooking Lake. Overall, manure provided the best residual effect at both sites.


Figure 4
View larger version (35K):
[in this window]
[in a new window]
  		Fig. 4. Residual effects of amendments on grain yield during 1991 through 1995 with different topsoil removals at two sites in Alberta.

 

   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
CONCLUSIONS
 REFERENCES
 
The results of this study showed marked effects of simulated erosion and amendments on grain yield at both sites in all years (Fig. 1 and 3, Tables 4 and 7). The drastic loss in crop productivity due to simulated erosion was associated with a marked reduction in the amount of nutrients and organic matter in the soil (Table 2) and N-mineralization potential (Table 3) after topsoil removal. Results from field research in the Prairie Provinces of Canada (Soper et al., 1971; Carson et al., 1974; Walker, 1975; Nyborg and Malhi, 1990) consistently showed an increase in crop-yield response to applied N when the level of extractable NO3–N (available N) in the soil decreased. Thus, yield reductions due to soil loss could also be explained as a function of nutrient losses. In our study, there was a close relationship between nutrient loss and yield reduction (Fig. 5 ). At both sites, the rate of yield reduction was a function of the mass of available N lost and either linear or curvilinear functions satisfactorily explained these relationships. In other words, the greater the loss in the ability of the soil to provide available N in the springtime, the greater the yield reduction.


Figure 5
View larger version (19K):
[in this window]
[in a new window]
  		Fig. 5. Effects of available and total nutrient loss from 5-, 10-, 15-, and 20-cm topsoil removal on wheat grain yield loss in 1991 at two sites in Alberta.

 
At Cooking Lake, the rate of yield reduction per unit of total N lost was independent of the mass of total N lost (i.e., the relationship between yield reduction and total N loss is linear). The effect of total N loss on yield reduction was greater at Cooking Lake than at Josephburg. In the same order, there was a much greater reduction in total N in the soil due to simulated erosion at Cooking Lake than at Josephburg (Table 2). For example, a loss of 500 kg ha–1 of total N in the upper 30-cm soil depth reduced wheat yield by 0.06 Mg ha–1 at Josephburg and by 0.42 Mg ha–1 at Cooking Lake. Had N losses been greater by an order of magnitude (i.e., 5000 kg ha–1), the respective yield reductions in comparison to uneroded soil profiles would have been 1.80 and 3.20 Mg ha–1.

There is usually a positive relationship found between crop yield and extractable P (Heapy et al., 1976; Nuttal et al., 1979; Malhi et al., 1992). Other researchers (Olsen et al., 1961; Leikam et al., 1983; Malhi et al., 1992) have also found that the effectiveness of fertilizer P increases with decreasing levels of extractable P. In our study, simulated erosion caused less drastic effects on extractable P at Josephburg than at Cooking Lake (Table 2). The curvilinear relationship of yield reduction and extractable P lost determined at Josephburg indicate that yield reductions would be rather small up to ~15 kg P ha–1 but increase substantially afterward (Fig. 5). This response at Cooking Lake was linear and reached higher levels of yield loss than those at Josephburg. The effect of total P loss on yield reduction was linear at both sites. Yield reductions as a function of total P at Cooking Lake would occur at a faster rate than at Josephburg (3.3 vs. 1.9 kg ha–1 of wheat grain yield lost for every 1 kg ha–1 of total P lost from the upper 30-cm soil depth).

Erosion may affect not only nutrient-pool sizes but also nutrient-release dynamics. In this study, erosion reduced not only the amount of available N (NO3–N; Table 2) but also the ability of the soil to release available N for crop uptake (Table 3). In turn, the amounts of N mineralized during the 100-d growing season, predicted at optimum soil water and 15°C from laboratory soil incubations, correlated strongly with plant N uptake under field conditions (R2 values of 0.93 at Josephburg and 1.00 at Cooking Lake). The correlation between N mineralization and N uptake was linear at Josephburg (y = 0.597x + 0.347) and quadratic at Cooking Lake (y = 0.097x2 – 0.138x + 1.355). An analysis of the shape of the curves suggests, however, that N uptake by plants growing in uneroded or moderately eroded soil at Cooking Lake was more efficient than at Josephburg (e.g., right half of both lines, Fig. 5). The slope of the line for Josephburg indicates that only 60% of the N theoretically mineralized during the 100-d growing season was taken up by the crop. In addition to inherent soil differences, we surmise that the use of N fertilizers, legumes, and cattle manure at Cooking Lake rendered N more easily mineralizable than that at Josephburg.

In earlier research, Larney et al. (1995a) described yield (y) as a function of simulated erosion (x) without any amendment added as: y = a – bx + cx2. This equation tells us that marginal yields diminish with the incremental loss of topsoil; i.e., the thickness of the topsoil near the surface influences yield more than topsoil quality. In our study, the exception was the equation for the soil at Josephburg that had also a negative c coefficient (Fig. 6 ). At Josephburg, therefore, the thicker topsoil mitigated the detrimental effect of erosion on grain yield.


Figure 6
View larger version (13K):
[in this window]
[in a new window]
  		Fig. 6. Effects of depth (cm) of simulated erosion (cut) and restorative amendments on wheat grain yield in 1991 at two sites in Alberta.

 
Results obtained in 1992 through 1995 generally supported the findings of 1991. They confirmed that soil productivity was a function of topsoil depth, which in turn was strongly a function of soil nutrients. These results were not repeated at Cooking Lake, where the soil has a much thinner A horizon. Wheat on 5-, 10-, 15-, and 20-cm cuts receiving chemical fertilizers yielded 107, 90, 81, and 75%, respectively, of wheat grown on the 0-cm cut without fertilizers. These results at Cooking Lake compare closely with those reported by Dormaar et al. (1986) in southern Alberta.

The yield–soil depth relationships discussed so far clearly attribute substantial weight to fertility–chemical soil properties as a means of explaining yield responses to simulated erosion. These results are in agreement with those reported by others (Dormaar et al., 1986; Shafiq et al., 1988; Tanaka and Aase, 1989; Verity and Anderson, 1990).

Other factors, however, such as increased bulk density and compaction can reduce root penetration and soil aeration (Sands et al., 1979) and therefore reduce crop yields (Tanaka and Aase, 1989). Results of root growth on selected treatments in 1991 indicated that root mass and length decreased dramatically by simulated erosion, at least at Cooking Lake, and the addition of amendments such as fertilizer and manure increased root growth. Soil bulk density, strength, or mechanical impedance has been reported to influence crop root growth and subsequent yield (Voorhees et al., 1975; Ehlers et al., 1983; Pierce et al., 1983; Blevins et al., 1984). The measured bulk densities in the surface soil ranged between 1.21 and 1.38 Mg m–3 for different depths in the eroded soil. Thus, bulk density may not have been a factor for root growth.

In this study, we did not observe soil compaction effects on germination and plant establishment. The reduced yield on eroded soil thus arose from low-tillering wheat plants with reduced LAI and lighter seeds rather than from low plant densities. Fertilizer increased yield by increasing tiller density, spike density, and LAI. Manure instead increased yield by increasing tiller density, spike density, LAI, and seed weight.

The application of fertilizer N and P and manure compensated the loss of yield even at the highest level of simulated erosion (Fig. 6). The return of 5 cm of topsoil to the plots also reduced the impact of yield loss due to erosion, but the improvement in yield was much less than with fertilizers or manure. In 1991, grain yield differences between the control and amendment treatments were much greater on the 20-cm cut than the 0-cm cut, indicating the importance of amendments in restoring crop productivity on severely eroded soils. In 1991, one season after the erosion treatment, the order of effectiveness of the various amendments tested was fertilizer > manure > topsoil, but in 1992 to 1995 the residual effectiveness of amendments treatments was in the order of manure > topsoil > fertilizer. This shows a greater and longer compensatory effect of manure in mitigating lost productivity due to topsoil erosion. On average of 5 yr, a one-time application of manure increased grain yield over the 20-cm eroded control by 129% at Josephburg and by 271% at Cooking Lake. Similar residual effects of manure, topsoil, and fertilizers on grain yield of wheat were obtained in southern Alberta, Canada, by Larney et al. (2000). Other researchers have reported the importance of manure application in increasing crop productivity and improving soil physical properties (Larney and Janzen, 1996, 1997; Arriaga and Lowery, 2003). The study of Larney and Janzen (1996) suggested that crop productivity on severely eroded soils can be restored with the addition of livestock manure and crop residues by substituting for lost topsoil.

From 1992 to 1995, the application of chemical fertilizers increased the grain yield of wheat considerably on eroded treatments, but maximum yields were attained in plots receiving application of manure along with chemical fertilizers. For example, at Josephburg in the 20-cm eroded treatment, the increase in grain yield over the unfertilized control was 0.31 Mg ha–1 with manure and 1.56 Mg ha–1 with manure + fertilizer. The corresponding values at Cooking Lake were 0.71 and 2.50 Mg ha–1. This indicated that, for high sustaining crop production, integrated use of chemical fertilizers and manure is needed because soil properties deteriorated by erosion can be improved with the application of organic amendments and chemical fertilizers in a short period (Izaurralde et al., 1998a).


   CONCLUSIONS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
REFERENCES
 
The incremental loss in crop productivity was a function of the depth of topsoil removed. Simulated erosion reduced yields more at Cooking Lake, with a 15-cm-thick A horizon, than at Josephburg, with a 30-cm-thick A horizon. Yield reductions due to simulated erosion were associated with the amounts of available and total N and P removed and a reduction in soil organic matter. Simulated erosion reduced not only the amounts of available N but also the ability of the soil to release mineral N for plant uptake. Laboratory predictions of N mineralization were closely correlated with plant N uptake in the field. There was apparently greater efficiency of N uptake by plants grown in either uneroded or moderately eroded soil at Cooking Lake than those grown at Josephburg. The three amendments tested to restore lost productivity varied in their effectiveness. In 1991, the season after the erosion treatment, the order in effectiveness was: fertilizer > manure > topsoil. But in 1992 and beyond, the residual effectiveness of amendments was manure > topsoil > fertilizer. The combined use of chemical fertilizers and manure produced the maximum grain yield at Cooking Lake. The results indicate the importance of an integrated use of organic amendments and chemical fertilizers for the best crop yields on severely eroded soils. These findings also suggest the need for further long-term research to determine the best sources and rates of organic amendments for sustainable high soil quality and productivity on severely eroded soils subject to any specific management system.


   ACKNOWLEDGMENTS
 
We acknowledge the excellent technical assistance of J. Brown, B. Hoar, M. Molina, J. DeMulder, C. Nguyen, J. Thurston, and Z. Zhang. This work was jointly supported by the Agriculture and Agri-Food Canada Greenhouse Gas Research Initiative and the Parkland Agriculture Research Initiative.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 





This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Izaurralde, R. C.
Right arrow Articles by Quiroga Jakas, M. C.