Biology and Fertility Biol Fertil Soils (1989) 8:134-143
of S o H s © Springer-Verlag1989
Shoot, root, soil and microbial nitrogen dynamics in two contrasting soils cropped to barley (Hordeum vulgate L.) P.M. Rutherford and N.G. Juma Department of Soil Science, University of Alberta, Edmonton, Alberta T6G 2E3, Canada
Summary. Dynamics
of barley N, mineral N, and organic N were compared at Ellerslie (Black Chernozem, Typic Cryoboroll) and Breton (Gray Luvisol, Typic Cryoboralf) in central Alberta, using 15N-urea. On average, shoot N and shoot 15N recoveries at Ellerslie (14.1 g m-2, 36%) were greater than at Breton (4.5 g m -2, 17%). Root N (g m -2) did not significantly differ between sites ( 0 - 30 cm) but root 15N recovery was greater at Breton (3.4%) than Ellerslie (1.8%). Low levels of shoot N and shoot 15N at Breton were partly due to very wet soil conditions in July, which resulted in premature shoot senescence and low plant N uptake. Although the total 15N recoveries from the system (to 30 cm depth) at Ellerslie (63%) and Breton (56%) were similar, soil 15N was greater at Breton (35%) than at Ellerslie (26o/o). There were no differences in mineral N between sites but the average ~SN recovery in the mineral-N pool was significantly greater at Ellerslie (3.3O/o) than at Breton (1.6%). There was no difference in tSN recovery in the microbial biomass ( - 3 %) between sites, although non-microbial organic ~SN was greater at Breton (31%) than at Ellerslie (20%). The two soils showed differences in the relative size of kinetically active N pools and in relative mineralization rates. Microbial N ( 0 - 3 0 cm) was greater at Ellerslie (13.3 g m -2) than at Breton (9.9 g m-Z), but total microbial N made up a larger proportion of total soil N at Breton (1.6%) than at Ellerslie (0.9%). In the 0 - 1 0 c m interval, microbial N was 1.7-fold greater and non-microbial active N was 3-fold greater at Breton compared to Ellerslie, when expressed as a proportion o f total soil N. Net N mineralization in a 10-day laboratory incubation was 1A-fold greater in the Black Chernozem ( 0 - 1 0 cm interval) from Ellerslie, compared to the Gray Luvisol from Breton, when ex-
Offprint requests to: N.G. Juma
pressed per gram of soil. Net N mineralization in the soil from Breton was double that of the soil from Ellerslie, when expressed as a proportion of soil N. Although soil N (g m -2) was 2.5-fold greater at Ellerslie compared to Breton, it was cycled more rapidly at Breton.
Key words: Active
N - Cryoboralf - Cryoboroll 15N - N cycling - H o r d e u m vulgate L. - Microbial biomass N
Soils belonging to the Chernozem and Luvisol orders are important for cereal production in Alberta and represent two extremes of natural fertility, in biological, physical, and chemical properties, and in crop productivity. Tracer studies have shown that up to 4 0 - 7 0 % of fertilizer N is taken up by annual crops, 1 0 - 3 0 % remains in the soil and the rest is unaccounted for (Stevenson 1986). The turnover of fertilizer N is affected by many factors, such as environmental conditions, cropping practices, placement, and timing (Olson and Kurtz 1982). Although the recovery of fertilizer N in harvested crops has been measured for a variety of conditions in these soils, information on the dynamics of N uptake by cereals over the growing season is scarce. More information is needed for specific agroecosystems throughout Alberta in order to develop more widely applicable concepts. A very small fraction of fertilizer N remains as NH~ and NO 3 in cropped soils at the end of the season (Stevenson 1986). Because soluble inorganic N is subject to crop uptake, leaching, denitrification, volatilization, immobilization, and clay fixation, it is important to determine the occurrence of major processes involved in partitioning of mineral N over the growing season.
135
Mineralization-immobilization turnover studies using classical fractionation, acid hydrolysis in various particle-size fractions and in whole soil, and the kinetics of biologically meaningful pools have been reviewed by Juma and McGill (1986). Mineralizat i o n - immobilization turnover is affected by the shape and location of substrates (McGill and Myers (1987), soil management (Campbell and Souster 1982), environmental conditions (Schimel and Parton 1986), texture and mineralogy (Monreal et al. 1981; van Veen et al. 1985), available energy: N ratio (Jansson 1958), and by interactions of the decomposer foodweb in soils (Coleman et al. 1984). There is a need to characterize the organic N into biologically meaningful fractions in order to effectively predict mineralization-immobilization turnover. Soil organic N can be kinetically separated into an active-N phase and a passive-N phase (Jansson 1958). The active-N phase is made up of the living microbial N and the non-microbial active N, consisting of microbial metabolites and/or recently stabilized microbial residues (Paul and Juma 1981). The non-microbial active N has a similar enrichment to that of microbial N in most isotope studies. Microbial N and non-microbial active N are the major sources and sinks of mineral N (Juma and Paul 1984). Active-N pools, estimated by isotope dilution techniques, account for 5-15°70 of soil N (Jansson 1958; Paul and Juma 1981). Knowledge of the size and activity of the active N pools in various soils is needed to predict soil N availability for crops over the growing season, but is in itself insufficient. The flux of N is the product of Npool sizes and their specific mineralization-rate constants. Consequently, a soil with a low organic-N content, but a high specific mineralization rate, may mineralize as much N as a soil with a high organic-N content and a low specific mineralization rate. Therefore, comparison of pool sizes between soils does not provide enough information to understand N dynamics. The objectives of this study were to determine (1) whether N dynamics and distribution and recovery of urea-15N in the barley-soil system of a Black Chernozem site differs from that of a Gray Luvisol site; and (2) whether the proportion of total N in kinetically active pools and the N flux through them differ in these soils.
Materials and methods Site description andsoils. This study was conducted at the University of Alberta Elleslie and Breton Research Plots. The Ellerslie Plots are located approximately 10kin south of Edmonton, Alberta (53 ° 25'N, 113 ° 33'W), on a Black Chernozem soil (Malmo series). The Breton Research Plots are located near the town of Breton, Alberta (53 ° 07' N, 114° 28' W), approximately 110 km southwest of
Table 1. Soil properties of Ellerslie and Breton Plots Bulk density Texture Depth Total C Total N pH (cm) (%) (%) ( l : 2 s o i l : H 2 0 ) (Mgm -3) (mass : volume)
Ellerslie (Black Chernozem; Typic Cryoboroll) 0 - 1 0 6.46 10-20 6.32 2o-3o 5.23
0.534 0.491 0.419
6.1 6.0 6.0
0.86 1.06 1.17
Silty clay loam
1.11 1.30 1.55
Silt loam
Breton (Gray Luvisol; Typic Cryoboralf)
0-10 2.17 10-20 1.92 20-30 1 . 1 3
0.185 0.164 0.104
6.2 6.3 6.1
Edmonton. The soil used was a Gray Luvisol (Breton series). Both soils have been in agricultural production for over 50 years. Detailed description of the sites and cropping histories have been described elsewhere (Rutherford and Juma 1989). Soil properties for both sites are presented in Table 1.
Experimental design and sampling schedule. The experiment was a factorial split-plot design with three blocks at two sites. Four openended cylinders (20 cm internal diameter, 30 cm long) were pressed into each block in late May 1986. On May 28 1986, the cylinders were seeded (eight seeds/cylinder) with barley (Hordeum vulgare L., cv Empress) and injected with 10 ml urea solution (23 mg N m l - l ; 75 kg N h a - l ; 5.2% 15N excess) at a depth of 2cm with a syringe. The area surrounding the cylinders was also seeded to barley and fertilized with urea (77 kg N ha -l) and triple superphosphate (47 kg P ha -s) 3 days before installation of the cylinders. No herbicides were used. At each sampling date the shoot material from a cylinder in each block was cut at ground level, then the cylinder was excavated from the block and the soil was divided into three depths, 0-10, 10-20, and 2 0 - 30 cm, in the laboratory. The roots were separated from the soil by hand. The soil samples were stored moist for determination of microbial biomass and mineral N. Subsamples were dried for total-N determinations. At Ellerslie the plots were sampled on July 31, August 18, September 8, and September 29, which corresponded to 64, 82, 103, and 124 days after sowing, respectively. At Breton the plots were sampled on August 11, September 1, September 22, and October 20, which corresponded to 75, 96, 117, and 145 days after sowing. No early-season samples were taken, due to wet conditions and laboratory renovations. The last sampling date at Breton was also delayed because of wet field conditions.
Analyses. The shoot and root material was dried at 70°C and weighed. The shoots were ground to 20 mesh in a Wiley mill and the root material was ground with a mortar and pestle. Total N was determined on soil, shoot, and root (0-10, 10-20 cm) samples by semimicroKjeldahl analysis with steam distillation after reduction of N O r with Fe (Bremner and Mulvaney 1982). Soil mineral N was determined on sieved soil (10 mesh) at the time of sampling. The soil sample was extracted in a 2 M KC1 mixture (25 g soil to 125 ml solution) for 1 h and was then filtered through Whatman 42 filter paper. The extractant was frozen until the time of analysis. Technicon AutoAnalyzers were used to determine ammonium N (Industrial Method 98-70W, 1973) and nitrate-N (Industrial Method 497-77A, 1977). Subsamples were treated with MgO, steam distilled, titrated, spiked with 0.5M HzSO4, and dried. Nitrate N was determined by reduction to ammonium-N with Devarda alloy followed by distillation. Microbial N (Bn) was determined by the method of Jenkinson and Powlson (1976) on moist, sieved (10 mesh) soil ajusted to 55o70
136 Table 2. Total N budget and barley N concentration and C : N ratio at Ellerslie and Breton (means of three replicates) Variable
Depth (cm)
Days after sowing a 64/75
82/96
103/117
124/145
13.3 2.50 14.9 1.1 0.2 1.04 1.24 17.5 462 534 462 1458
12.0 2.08 20.5 0.5 0.1 1.30 1.13 23.6 450 535 477 1462
18.9 2.14 19.7 0.7 0.2 1.56 1.42 19.2 462 539 515 1516
12.0 1.79 23.7 0.9 0.2 1.47 1.76 21.8 448 482 502 1432
4.7 1.15 33.5 0.9 0.2 1.24 1.62 19.2 203 211 149 563
3.6 1.37 25.2 0.8 0.1 1.56 1.32 19.2 197 218 146 561
Ellerslie Shoot N (g m -2) Shoot N (°7o) Shoot C" N ratio d Root N e (g m - 2 )
0 - 10 10-20 0 - 10 10-20 0 - 10 0 - 10 1 0 - 20 2 0 - 30
Root N e (070) Root C : N ratio d Soil N (g m -2)
Total soil N
Breton Shoot N (g m -2) Shoot N (°70) Shoot C : N ratio d Root N e ( g m -2)
4.3 0.71 37.6 0.8 0.1 1.24 1.45 21.1 195 206 140 541
0-10 1 0 - 20 0 - 10 10-20 0-10 0 - 10 1 0 - 20 2 0 - 30
Root N e (07o) Root C : N ratio a Soil N (g m -2)
Total soil N Source of variation
df
Shoot N
Mean square of analysis of variance c Site 1 546" Error 1 4 49.9 Date 3 18.0" Site×date 3 15.7" Error 2 12 3.0 Depth 2 Site×depth 2 Error 3 8 Date × depth 6 Site × date x depth 6 Error 4 24
5.4 1.21 30.3 0.9 0.03 b 1.28 0.47 b 24.0 228 214 206 648
Shoot N (o70)
Shoot C : N
Root N
Root N (%)
Root C : N
Soil N
6.21 *** 0.09 0.007 0.49** 0.04
854** 31.9 5.6 117" 12.4
0.007 0.027 0.042 0.099 0.056 5.22*** 0.030* 0.002 0.019 0.085 0.002
0.09 0.16 0.42 0.13 0.11 0.01 0.10 0.02 0.20 0.16 0.07
0.78 2.03 27.6* 9.50 2.03
1.6 × 10s *** 3909 1717 1780 3 083 12811 10280 4391 619 1503 2500
a Sampling date for Ellerslie and Breton, respectively bn=l rather t h a n n = 3 c The difference between m e a n s is significant at, * P < 0 . 0 5 ; * * P < 0 . 0 1 ; ***P_<0.001 a Shoot and root C have been reported by Dinwoodie a n d J n m a (1988) e Since root N was measured for two depths, the degrees o f freedom are depth (1), s i t e × d e p t h (1), error 3 (4), d a t e × d e p t h (3), site × date x depth (3) a n d error 4 (12)
water-holding capacity, using the equation B n = Fn/Kn, where F n is {[(mineral N (NH~ +NO~-) in fumigated soil incubated for 10 days] - [mineral N in non-fumigated soil incubated for 10 days]} and K n = 0.68 (Shen et al. 1984). The change in mineral N during the incubation o f the non-fumigated samples was used as a measure o f net N mineralization. For the isotope ratio analysis of various N fractions, N H ~ in titrated samples was oxidized to N 2 using LiOBr (Porter and O ' D e e n 1977) a n d the isotope analysis was determined using a Micromass 602C double-collection mass spectrometer. Microbial 15N excess
was a s s u m e d to be the same as the lSN excess of N H ~ extracted from the fumigated samples after a 10-day incubation. Non-microbial organic 15N was calculated by subtracting the s u m of microbial lSN and mineral 15N from total soil 15N. Non-microbial active 14N (NMAN) was calculated by the m e t h o d of Paul a n d J u m a (1981): N M A N = non-microbial organic lSN/atom % excess of microbial N.
Statistical analyses. A n analysis of variance was performed on all experimental variables, using UANOVA, a multivariate analysis o f
137 covariance program developed at the University of Alberta (T. Taerum, personal communication 1987). The procedure gives probabilities which are adjusted for unequal variances of the m e a n s (Greenhouse and Geisser 1959).
Results
Shoot, root and soil N trends Shoot N (Table 2) was, on a average, 3.1-fold greater at Ellerslie than at Breton, peaking at the third sampiing date at Ellerslie and on the second sampling date at Breton. The average shoot-N concentration was 1.9-fold greater at Ellerslie than at Breton and decreased with time at EUerslie but increased at Breton.
The average shoot C : N ratios (Table 2) were lower at Ellerslie (19.7) compared to Breton (31.7) and showed opposite trends over time at the two sites. Root N decreased with depth but the decrease was greater at Breton than at Ellerslie. The root-N concentration (Table 2) did not differ significantly between sites, dates, or depths. The root C : N ratios (Table 2) did not differ significantly between sites in the 0 - 1 0 cm interval, but ratios peaked on the second sampling date at both sites. Total soil N (Table 2) was 2.5-fold greater at Ellerslie compared to Breton. The N concentration decreased with depth (Table 1) but total soil N (g m -z) did not decrease with depth because bulk densitiy increased with depth (Table 1). Mineral N trends
Table 3. Mineral N (g m -2) at Ellerslie and Breton (means of three replicates) N pool
Ellerslie Ammonium N
Nitrate N
Mineral N
Breton Ammonium N
Nitrate N
Mineral N
Depth (cm)
Days after sowing a 64/75
82/96
103/117
124/145
0 - 10 1 0 - 20 2 0 - 30 0-10 10-20 20 - 3 0 0-10 10-20 20-30
0.34 0.40 0.67 0.79 0.80 2.87 1.12 1.20 3.54
0.33 0.32 0.63 0.46 0.52 1.49 0.79 0.84 2.12
0.44 0.27 0.70 0.87 0.62 1.55 1.31 0.90 2.25
0.16 0.18 0.54 0.52 0.72 1.02 0.68 0.89 1.56
0-10 10-20 2 0 - 30 0-10 10-20 2 0 - 30 0 - 10 10-20 2 0 - 30
0.29 0.49 0.76 0.90 0.72 0.49 1.19 1.21 1.24
0.85 0.33 0.55 0.54 0.48 0.48 0.89 0.81 1.03
0.17 0.22 0.39 0.52 0.69 0.53 0.68 0.91 0.93
0.15 0.33 0.40 0.43 0.41 0.33 0.57 0.74 0,73
Nitrate N
Mineral N
4.09 1.04 0.94* 0.20 0.16 1.77 2.87 0.63 0.19 0.38 0.08
4.89 0.87 1.65" 0.18 0.26 4.55" 3.37 0.61 0.24 0.31 0.15
Source of variation
df
Ammonium N
Mean square of analysis of variance b Site 1 0.036 Error 1 4 0.015 Date 3 0.13" Site x date 3 0.059 Error 2 12 0.026 Depth 2 0.65"** Site×depth 2 0.043 Error 3 8 0.009 Datexdepth 6 0.015 S i t e x d a t e x depth 6 0.010 Error 4 24 0.020
Ammonium N decreased with time and increased with depth at both sites (Table 3). Nitrate N and total mineral N were greatest in the 2 0 - 3 0 c m interval at
Table 4. Microbial N and net N mineralized ( g m -2) from nonfumigated soil samples during 10-day incubation (means of three replicates) N pool
Ellerslie Microbial N
Net N mineralized Breton Microbial N
Net N mineralized
Depth (cm)
64/75
82/96
103/117
124/145
0-10 10 - 20 20 - 30 0 - 10 1 0 - 20 2 0 - 30
6.52 5.75 3.99 0.87 0.64 0.70
4.49 5.37 2.29 0.92 0.66 0.47
5.10 3.73 2.01 1.09 0.67 0.40
5,72 5.08 3.02 0.82 0.37 0.63
0 - 10 10-20 2 0 - 30 0-10 10-20 20-30
3.49 2.98 1.80 0.78 0.42 0.12
5.47 4.57 3.24 0.88 0.51 0.39
3.41 2.83 1.36 0.95 0.47 0.14
4.23 3.27 1.73 0.66 0.36 0.17
Source of variation
a Sampling date for Ellerslie and Breton, respectively b The difference between means is significant at *P_<0.05; **P_<0.01; ***P_< 0.001
Days after sowing a
df
Microbial N
Mean square of analysis of variance b Site 1 26.9* Error 1 4 3.29 Date 3 4.84 Site x date 3 7.14 * Error 2 12 1.01 Depth 2 36.5*** Site × depth 2 0.92 Error 3 8 0.98 Date x depth 6 0.34 Site x date x depth 6 0.49 Error 4 24 24.0 a,b See footnotes to Table 3
Net N mineralized
0.71 0.11 0.065 0.033 0.047 1.56"** 0.099 0.028 0.054 0.030 0.029
138
Ellerslie but were similar for all depths at Breton. Nitrate-N decreased on sampling date 2, increased on date 3 and decreased on the last sampling date at both sites.
Microbial N trends and net N mineralization in the laboratory Microbial N (Table 4) was on average 1.4-fold greater at Ellerslie than at Breton. Microbial N declined over the first three dates at Ellerslie and increased on the last date; in contrast, it peaked at Breton on date 2. Microbial N decreased with depth at both sites.
Table 5. Percentage 15N recovery in the plant-soil system at Ellerslie and Breton (means of three replicates) N pool
Depth
Net N mineralization was 1.4-fold greater in the soil from Ellerslie when expressed per gram soil (data not shown), but was not significantly different between the two soils when expressed on an area basis (g m -2) (Table 4). Net mineralization ( g m -2) generally declined with depth at both sites.
15N recovery in the plant-soil system Shoot 15N was the only 15N pool measured that changed significantly between sampling dates (Table 5). Shoot-aSN recoveries were greater at Ellerslie
Table 6. Percentage 15N recovery in soil N pools at Ellerslie and Breton (means of three replicates) N pool
Days after sowing a
Depth (cm)
64/75
82/96
103/117
124/145
0 - 10 1 0 - 20 20 - 30 0 - 10 10-20 2 0 - 30 0 - 10 10-20 20-30
1.28 1.67 2.04 3.14 0.55 0.26 14.56 1.22 0.12
1.28 0.70 1.40 1.82 0.45 0.23 19.08 4.95 -0.52
1.60 0.58 0.93 2.29 0.29 0.16 16.70 4.27 0.31
0.40 0.57 0.80 2.08 0.30 0.22 13.95 3.53 0.18
O - 10 10-20 2 0 - 30 O - 10 10 - 20 2 0 - 30 O - I0 10-20 20 - 30
1.24 0.69 0.67 2.15 1.24 0.60 18.65 4.89 2.05
0.72 0.40 0.19 1.77 0.97 0.35 27.72 6.31 2.93
1.02 0.58 0.33 1.66 0.92 0.24 29.65 4.62 1.47
0.30 0.17 0.24 1.36 0.44 0.23 20.97 3.00 0.86
(cm) Ellerslie Shoot N Root N c Total soil N
0 - 10 10 - 20 0-10 1 0 - 20 20-30
Total recovery Breton Shoot N Root N ¢ Total soil N
0-10 10-20 0-10 10 - 20 20-30
Total recovery Source of variation
df
64/75
82/96
103/117
124/145
41.4 1.8 0.3 19.0 3.4 2.4
33.9 1.2 0.2 22.2 6.1 1.1
40.7 1.6 0.3 20.6 5.1 1.4
26.4 1.5 0.3 16.4 4.4 1.2
24.8 68.3
29.4 64.7
27.1 69.7
22.0 50.2
20.5 3.5 0.5 21.7 7.0 3.4
19.7 3.3 0.1 30.2 7.7 3.5
17.9 3.1 0.4 32.3 6.3 1.9
11.3 2.4 0.2 22.6
32.1 56.6
41.4 64.5
40.5 61.9
27.5 41.4
Shoot 15N recovery
Mean square of analysis of variance b Site 1 1992" Error 1 4 172 Date 3 172' Site x date 3 26.9 Error 2 12 27.3 Depth 2 Site x depth 2 Error 3 8 Date x depth 6 Site x date x depth 6 Error 4 24
Root lSN recovery
3.6
1.3
Ellerslie Total mineral N
Microbial N
Non-microbial organic N Breton Total mineral N
Microbial N
Non-microbial organic N
Soil 15N recovery Source of variation
10.4"* 0.19 0.65 0.33 0.43 42.4*** 6.21 *** 0.05 0.21 0.15 0.24
149"* 6.4 55.3 6.9 14.0 3015"** 78.2 * 8.3 21.3 10.0 8.3
a Sampling date for Ellerslie and Breton, respectively b The difference between means is significant at, * P < 0 . 0 5 ; * * P < 0 . 0 1 ; ***P--< 0.001 c Since root N was measured for two depths, the degrees of freedom are depth (1), site x depth (1), error 3 (4), date x depth (3), site x date x depth (3) and error 4 (12)
Days after sowing a
df
Total mineral 15N recovery
Mean square of analysis of variance b Site 1 5.56" Error 1 4 0.36 Date 3 2.19 Site x date 3 0.21 Error 2 12 0.66 Depth 2 0.59 Site x depth 2 0.64 Error 3 8 0.35 Date x depth 6 0.28 Site × date × depth 6 0.20 Error 4 24 0.39 a,b See footnotes to Table 3
Microbial lSN recovery
Nonmicrobial organic 15N recovery
0.003 0.78 1.01 0.12 0.28 20.4*** 1.87" 0.13 0.21 0.12 0.26
235 ** 7.3 63.5 6.5 16.3 2483 *** 93.5** 7.0 17.6 10.2 6. i
139 Table 7. Microbial N and mineral N expressed as a proportion of soil N and microbial 15N a n d mineral 15N expressed as a proportion of soil 15N at Ellerslie and Breton (means of three replicates) N pool
Depth (cm)
Days after sowing a 64/75
82/96
103/117
124/145
Ellerslie % of soil N Microbial N
Mineral N
0-10 1 0 - 20 2 0 - 30 0 - 10 1 0 - 20 2 0 - 30
1.40 1.07 0.86 0.25 0.23 0.77
1.00 1.12 0.48 0.18 0.16 0.42
1.11 0.70 0.40 0.28 0.17 0.43
1.28 1.08 0.55 0.15 0.19 0.31
°/0 of soil 15N Microbial 15N
Mineral 15N
0 - 10 10-20 2 0 - 30 0-10 1 0 - 20 2 0 - 30
16.6 16.2 14.1 6.9 45.8 70.9
8.2 11.8 3.8 6.0 21.2 22.4
11.5 7.6 14.1 7.9 12.9 64.3
13.0 7.9 17.0 2.6 15.7 71.3
Breton % of soil N Microbial N
Mineral N
0-10 10 - 20 20 - 30 0 - 10 10-20 2 0 - 30
1.78
2.39
1.68
2.14
1.57
2.13
1.34
1.51
1.14 0.60 0.58 0.94
1.51 0.39 0.37 0.58
0.88 0.34 0.43 0.66
1.13 0.29 0.34 0.59
% of soil 15N Microbial 15N
Mineral 15N
Source of variation
0 - 10 10-20 2 0 - 30 0 - 10 10-20 20-30 df
Mean square of analysis of variance b Site 1 Error 1 4 Date 3 Site × date 3 Error 2 12 Depth 2 Site x depth 2 Error 3 8 Date x depth 6 Site x date × depth 6 Error 4 24 a,b See footnotes to Table 3
10.3 17.8 16.9 4.5 10.1 21.3
6.0 10.8 8.5 2.4 5.6 6.2
5.0 13.6 12.7 3.1 8.7 17.3
5.9 12.8 12.3 1.3 4.7 16.7
Microbial N: Soil N
Microbial 15N: Soil 15N
Mineral N: Soil N
Mineral 15N: Soil 15N
8.57 ** 0.23 0.52** 0.53 ** 0.030 3.24 *** 0.066 0.042 0.044 0.029 0.047
9.63 155 165 6.24 82.6 55.4 107 59.2 29.3 22.3 51.9
0.83 ** 0.025 0.22** 0.013 0.025 0.62 ** 0.005 0.031 0.026 0.014 0.020
7250" 533 808 250 275 6252 * 2292 372 513 232 127
140
(36%) compared to Breton (17%). At both sites shoot-15N recoveries declined, so that recoveries on date 4 were much lower than those on date 1. Shoot ~SN was the largest pool at Ellerslie but total soil ~SN was the largest pool at Breton. On average, root ~SN was 1.9-fold greater and soil 15N was 1.4-fold greater at Breton compared to Ellerslie. Root ~SN decreased more with depth at Breton than at Ellerslie. Soil 15N decreased with depth at both sites but declined more at Breton. On average, the total recoveries of applied urea-~SN in the soil-plant system were greater at Ellerslie (6307o) than at Breton (56%). Distribution of 15N in soil fractions lSN recovery as mineral N (Table 6) was 1.6 to 2.6-fold greater at Ellerslie than at Breton. There was no difference in ~SN recovery in the microbial biomass between sites. Microbial 15N decreased with depth at both sites but declined more at Ellerslie. The largest pool of ~SN in the upper two depths was non-microbial organic N (Table 6). This pool was greater at Breton (31%) than at Ellerslie (20%). The non-microbial organic 15N recovery decreased with depth at both sites but the decline was less with depth at Ellerslie.
Microbial and mineral N as a proportion of soil N In contrast to the size of the microbial-N pool (g m -2) (Table 3), microbial N expressed as a proportion of soil N (Table 7) was on average 1.7-fold greater at Breton than at Ellerslie. Over all depths at Ellerslie, microbial N accounted for 0.9°70 of total soil N but contained 11.8070 of 15N applied. At Breton microbial N accounted for 1.6070 of total soil N and contained 11.1070 o f 15N. Microbial N expressed as a proportion of soil N (%0) changed with date and decreased with depth. Microbial tSN, expressed as a proportion of soil 15N did not differ significantly with respect to site, date, or depth. Mineral N was greater at Breton than at Ellerslie when expressed as a proportion of soil N (Table 7). Mineral N (%0 of soil N) varied with date and increased with depth. A large proportion of soil ~SN was present as mineral N at both sites despite the small size o f the mineral-N pool (Table 7). A larger proportion of soil 15N was present as mineral N at Ellerslie than at Breton. The proportion o f mineral ~SN to soil 15N increased with depth, especially at Ellerslie where mineral 15N accounted for about 57070 o f soil 15N in the 2 0 - 3 0 cm interval. Absolute versus proportional N pool sizes and net N mineralization in 0 - 1 0 cm soil samples
Table 8. Absolute versus proportional N pool sizes and net N mineralized in 0 - 10 cm soil from Ellerslie and Breton (means of 12 replicates) Variable
Site
Significance of site effect a
Ellerslie Breton
Absolute (per m 2 or g soil basis) Soil N (g m -2) Microbial N (g m - 2 ) N M A N b (gin -2) Total mineral N ( g m -2) Net N mineralized (gg g-1 soil 10 days -1) Net 15N mineralized (gg 100 g - i soil 10 d a y s - l )
456 5.5 46.1 0.98
206 4.1 66.2 0.83
*** NS NS NS
10.4
7.2
*
4.2
3.0
NS
Proportional (per unit N basis) Microbial N (070 of soil N) NMAN b (070 o f soil N) Total mineral N (070 o f soil N) Net N mineralized (mg g-1 microbial N 10 days l) Net N mineralized (mg g N M A N b 10 days -1) Net N mineralized (rag g-1 soil N 10 days -1) Net lSN mineralized (°7o of soil 15N l0 days -1)
1.2 10 0.21
2.0 ** 32 ** 0.40 **
17
20
NS
Soil N (g m -2) and net N mineralization (~tg g-1 soil 10 days -z) were significantly greater in soil samples from Ellerslie than from Breton in the 0 - 1 0 cm interval (Table 8). The only variable that showed a significant site by date interaction (Table 8) was microbial N (%0 of soil N; P_<0.007). Microbial N (gm-Z), nonmicrobial active N (g m-Z), total mineral N (g m -2) and net I5N mineralization (~tg g - t soil 10 days-)) were not significantly different between the two soils. Microbial N, non-microbial active N and mineral N made up a larger proportion of soil N at Breton compared to Ellerslie. Net N mineralization was not different between soils when expressed as a proportion of microbial N, but was greater in soil from Ellerslie when expressed as a proportion of non-microbial active N. Net N mineralization was greater in soil from Breton when expressed as a proportion of total soil N. Net 15N mineralization did not significantly differ between soils when expressed as a proportion o f soil 15N.
2.4
1.5
*
0.20
0.39 **
Discussion
4.8
3.2
N dynamics in the two agroecosystems
NS
a The difference between means is significant at *P_<0.05; **P_<0.01; ***P<0.001; NS, not significant b Non-microbial active N
Differences in environmental conditions and soil properties at the two sites resulted in different N dynamics and a different distribution and recovery of urea-~SN.
141
Rainfall was probably the major environmental factor, because Breton received approximately 285 mm in July, almost three times the 30-year mean of 95 mm, while Ellerslie received 126mm, approximately 1.5 times the 30-year mean of 85 mm. Wet conditions affected plant growth at Breton, causing premature senescence. In July the plants at Breton were pale green with a shoot-N concentration of 0.71%o, which is considered to be deficient (Engel and Zubriski 1982; Olson and Kurtz 1982). Shoot-C to root-C ratios at Breton were, on average, 60% of those at Ellerslie (Dinwoodie and Juma 1988) and may reflect the more favorable N status at Ellerslie. Low shoot-15N recoveries at Breton compared to Ellerslie were due to less early season fertilizer uptake or to a greater loss of shoot ~SN through senescence (Sallam and Scott 1987). The higher root 15N at Breton was probably due to reduced N translocation to the shoots during the wet conditions in July. The shoot-N concentration declined as the plants matured at Ellerslie, but increased at Breton as the plants recovered from the wet conditions, perhaps having conserved N during their premature senescence (Trought and Drew 1980). The large decline in shoot 15N on sampling date 4 at both sites indicates a significant loss of N in the fall. The availability of N to cereals is affected by dynamics within the mineral pools. Leaching losses occurred at both sites. Soil samples taken from below the cylinders ( > 3 0 c m ) on sampling date 1 (data not shown) showed mineral 15N abundances in excess of the natural abundance at both sites. The decline of nitrate N with depth at Breton was probably due to early-season deep leaching accompanied by denitrification. The percentage water-filled porosity was very high ( > 8 0 % ) all season in the 2 0 - 3 0 c m interval (Rutherford and Juma 1989) and the nitrate N would have been prone to denitrification (Linn and Doran 1984) if C was available. Monreal and McGill (1983) found that denitrification at Breton may be restricted by a low supply of water-soluble organic C. Dinwoodie and Juma (1988) found a low level of water-soluble organic C at Breton; however, a small pool size does not necessarily imply a low level of supply. The greater recovery of mineral 15N at Ellerslie despite a non-significant site effect in mineral N indicates that there was less mineralization of immobilized 15N or perhaps there were greater leaching and denitrification losses at Breton. Similar trends between the mineral-15N and nitrate-N distribution indicate that most mineral 15N occurred as nitrate. The decline in mineral 15N between sampling dates 3 and 4 at both sites was probably due to leaching and/or denitrification since rainfall occurred during this time. Despite greater absolute mineral N (gm -2) at Ellerslie, a larger proportion of soil N was present as
mineral N at Breton. Thus net N mineralization was proportionally greater, or N losses and plant uptake of mineral N were proportionally less at Breton. At both sites the recovery of soil 15N in the mineral pool increased with depth as mineral N was leached into the profile. The much greater proportion of soil ~SN as mineral N in the 2 0 - 3 0 cm interval at Ellerslie compared to Breton was probably due to greater N losses at Breton. The availability of mineral N is affected by mineralization and immobilization reactions mediated by the microbial biomass. Microbial N (gm -2) was greater at Ellerslie because the soil had a much greater C and N content. This was expected, since microbial biomass has been shown to positively correlate with total organic C and N in soil (Schntirer et al. 1985; Carter 1986). Microbial ~SN did not differ between sites, even though microbial N was greater at Ellerslie, indicating a greater ~SN enrichment at Breton. The decline over time in microbial ~SN at both sites, though not statistically significant, was due to dilution of the microbial N pool with non-labelled N as the microbial biomass turned over. The large recovery of 15N in the microbial pool at both sites ( 4 - 1 8 % of soil 15N) indicates the importance of the microbial biomass as a significant sink and source of mineral N. The large recovery of ~SN in the non-microbial organic-N pool at Breton may be due to inherent differences in soil properties between the sites. BergstrOm (1986) found that the minerai-N supply decreased to prefertilization levels within 3 weeks of initial N uptake by barley. In this study urea hydrolysis would have been essentially complete after 3 - 4 weeks, and most fertilizer 15N should have been partitioned between mineral N, microbial N, and barley N before the heavy rains began in July. Thus the large recovery of 15N in the non-microbial organic pool was due to greater early-season immobilization at Breton. In addition, mineralization is much slower under anaerobic conditions (Patrick 1982) and wet conditions at Breton may have slowed the conversion of organic ~SN into mineral forms. However, anaerobic conditions in the 0 - 1 0 cm interval at Breton only existed for a short time and it is doubtful that the legacy of these conditions lasted until late summer. Perhaps early-season wet conditions caused an increase in available C through root death, increased root exudation (Dinwoodie and Juma 1988), microbial fermentation, or microbial death. Upon drying, re-immobilization of mineral 15N into the soil microbial biomass and the non-microbial organic pool would occur. This is consistent with the large soil-15N recovery and low shoot-15N recovery at Breton. Recoveries of 15N in the non-microbial organic pool at Breton in 1984 under oats and alfalfa were similar to those in 1986 despite drier field conditions
142 and the addition of fertilizer 15N when the oat plants were 1 week old and alfalfa was in its second year (Juma, personal communication). Thus immobilization of fertilizer N occurs rapidly in the first 4 weeks after application.
Relationships Between N, C, microbial, and faunal activities The two soils showed differences in the relative size o f the kinetically active N pools and in relative mineralization rates. The microbial biomass and non-microbial active N pools are major sources and sinks of mineral N in soil. Though total N was much lower at Breton compared to Ellerslie, 34% of soil N in the 0 - 1 0 cm samples was in the microbial and the nonmicrobial active-N pools at Breton compared to 12% at Ellerslie. The results of Campbell and Souster (1982) are comparable to those of the present study. They found that potentially mineralizable N accounted for a larger proportion o f soil N in cultivated Gray Luvisols compared to Black Chernozems. Turnover rates were not determined for the two pools, although there was no significant difference in net N mineralization between sites when expressed as a proportion of microbial N. Net N mineralization was less at Breton when expressed as a proportion of non-microbial active N. However, the proportionally large size of the active pools at Breton and the greater activity of the microbial biomass resulted in greater net N mineralization expressed as a proportion of total N at Breton compared to Ellerslie. This may have contributed to the greater proportion of soil N present as mineral N at Breton during the summer of 1986. The greater proportional C activity at Breton may be related to a relatively more active food web at Breton compared to Ellerslie (Rutherford and J u m a 1989). A greater proportion of soil C was respired in the Gray Luvisol compared to the Black Chernozem in a 10-day laboratory incubation. Microbial C and most faunal counts were greater at Ellerslie when expressed per square meter, microbial C and protozoa were greater at Breton and nematodes and microarthropods were not significantly different when expressed relative to soil C. Thus the C at Breton was proportionally more available to support biological activity than the C at Ellerslie. Rutherford and J u m a (1989) have proposed a hypothesis that in 1986 the food web was more active at Breton compared to Ellerslie. Differences in the detrital trophic web structure affect C and N turnover in the soil and the size and activity of the active-N pools. High C activity and relatively large decomposer food web kept a large proportion of soil N in the soil organisms at Breton. A large active microbial-N pool would be expected to release a large quantity of microbial
metabolites and by-products into the soil. Consequently, the non-microbial active-N pool would be large. Thus, a rapid C turnover and greater microbial and faunal activity resulted in a greater proportional net N-mineralization rate at Breton. The greater activity of C and N at Breton may have been due to differences in the quality of organic matter, mineralogy, soil texture, and soil organism activity at the two sites. A large proportion of the organic matter at Ellerslie may be stabilized and unavailable for rapid microbial decomposition. Pawluk (1986) hypothesized that the soil organic matter was resilient at Ellerslie because the soil had a high smectite content in the A horizon. This stabilizing influence may not have been as strong in the Ap at Breton, due to its lower clay and smectite content (Pawluk 1980). The decomposer population is protected in fine-textured soils (van Veen et al. 1984). A fine soil texture reduces the mineralization of C and N from plant residues (Ladd et al. 1981; Ladd et al. 1985), the turnover of C and N through the microbial biomass (Merckx et al. 1985; van Veen et al. 1985), and the proportion of soil N which is mineralizable (Campbell and Souster 1982).
Implications The present study showed that N dynamics in the Chernozem and the Luvisol soils were markedly different. Since N d y n a m i c s are affected by physical, chemical, and biological properties and associated processes, innovative m a n a g e m e n t techniques are needed in order to conserve C and N in soils and to increase the N uptake by crops. This study also showed that a comparison of N-pool sizes only yields partial information on N dynamics. The comparison of flux between different N pools is important, and should be considered when comparing nutrient dynamics in various soils and in evaluating different management practices.
Acknowledgments. We thank NSERC for financial support; Alberta Agriculture for the use of the coring truck; L. Toerper, J. Khatkar, C. Slupsky, C. Nguyen, J. Konwicki, C. Figueiredo, and J. Thurston for technical assistance; Dr R. T. Hardin and Dr T. Taerum for guidance on statistical analysis; G. Dinwoodie for valuable ideas and help on many aspects of the project; and Dr W.B. McGiU and Dr G.R. Webster for reviewing the manuscript.
References Bergstr0m L (1986) Distribution and temporal changes of mineral N in soils supporting annual and perennial crops. Swed J Agric Res 16:105-112 Bremner JM, Mulvaney CS (1982) Nitrogen-total. In: Page AL, Miller RH, KeeneyDR (eds) Methods of soil analysis: II. Chemical and microbiological properties, 2nd edn. Agronomy 9, Am Soc Agron, Madison, Wisc, pp 595-624
143 Campbell CA, Souster W (1982) Loss of organic matter and potentially mineralizable nitrogen from Saskatchewan soils due to cropping. Can J Soil Sci 62:651-656 Carter MR (1986) Microbial biomass as an indicator for tillage-induced changes in soil biological properties. Soil Tillage Res 7:29-40 Coleman DC, Anderson RV, Cole CV, McClellan JF, Woods LE, Trofymow JA, Elliott ET (1984) Roles of protozoa and nematodes in nutrient cycling. In: Todd RC, Giddens JE (eds) Microbial-plant interactions. Spec Publ 47, Am Soc Agron, Madison, Wise, pp 17-28 Dinwoodie GD, Juma NG (1988) Allocation and microbial utilization of C in two soils cropped to barley. Can J Soil Sci 68:495 - 505 Engel RE, Zubriski JC (1982) Nitrogen concentrations in spring wheat at several growth stages. Commun Soil Sci Plant Anal 13:531-544 Greenhouse SW, Geisser S (1959) On the methods in the analysis of profile data. Psychometrika 24:95 - 112 Jansson SV (1958) Tracer studies on nitrogen transformations in soil. Ann R Agric Coll Sweden 24:101-361 Jenkinson DS, Powlson DS (1976) The effects of biocidal treatments on metabolism in soil: V. A method for measuring soil biomass. Soil Biol Bioehem 8:209-213 Juma NG, McGill WB (1986) Decomposition and nutrient cycling in agro-ecosystems. In: Mitchell M J, Nakas JP (eds) Micro floral and faunal interactions in natural and agro-ecosystems. Nijhoff/ Junk, Dordrecht, pp 74-136 Juma NG, Paul EA (1984) Mineralizable nitrogen: Amounts and extractability ratios. Soil Sci Soc Am J 48:76-80 Ladd JN, Oades JM, Amato M (1981) Microbial biomass formed from 14C-, aSN-labelled plant material decomposing in soils in the field. Soil Biol Biochem 13:119 - 126 Ladd JN, Amato M, Oades JM (1985) Decomposition of plant material in Australian soils: III. Residual organic and microbial biomass C and N from isotope-labelled legume material, decomposing under field conditions. Aust J Soil Res 23:603-611 Linn DM, Doran JW (1984) Effect of water-filled pore space on carbon dioxide and nitrous oxide production in tilled and nontilled soils. Soil Sci Soc Am J 48:1267-1272 McGill WB, Myers RJK (1987) Controls on dynamics of soil and fertilizer N. In: Follett R (ed) Soil Fertility and organic matter as critical components of production systems. Spec Publ 19, Soil Sci Soc Am and Am Soc Agron, Madison, Wisc, pp 7 3 - 9 9 Merckx R, den Hartog A, van Veen JA (1985) Turnover of root-derived material and related microbial biomass formation in soils of different texture. Soil Biol Biochem 17:565-569 Monreal M, McGill WB (1983) Soluble C use during denitrification in saturated soil samples. Proceedings of 20th Annual Alberta Soil Sci Workshop, Feb 22, 23 1983, Edmonton, Alberta, pp 45 - 49
Monreal CM, McGill WB, Etchevers JD (1981) Internal nitrogen cycling compared in surface samples of an Andept and a Mollisol. Soil Biol Biochem 13:451-454 Olson RA, Kurtz LT (1982) Crop nitrogen requirements, utilization and fertilization. In: Stevenson FJ (ed) Nitrogen in agricultural soils. Agronomy 22. Am Soc Agron, Madison, Wise, pp 567-604 Patrick WH Jr (1982) Nitrogen transformations in submerged soils. In: Stevenson FJ (ed) Nitrogen in agricultural soils. Agronomy 22, Am Soc Agron, Madison, Wise, pp 449-465 Paul EA, Juma NG (1981) Mineralization and immobilization of soil nitrogen by soil microorganisms. In: Clark FE, Rosswall T (eds) Terrestrial nitrogen cycles. Ecol Bull (Stockh) 33:179-195 Pawluk S (1980) Micromorphological investigations of cultivated Gray Luvisols under different management practices. Can J Soil Sci 60:731-745 Pawluk S (1986) Vegetation and management effects upon some soil properties of Black Chemozemic soils of the Edmonton region. Can J Soil Sci 66:73-89 Porter LK, O'Deen WA (1977) Apparatus for preparing nitrogen from ammonium chloride for nitrogen-15 determinations. Anal Chem 49:514-516 Rutherford PM, Juma NG (1989) Dynamics of microbial biomass and soil fauna in two contrasting soils cropped to barley. Biol Fertil Soils 8:144-153 Sallam A, Scott HD (1987) Effects of prolonged flooding on soybeans during early vegetative growth. Soil Sci 144:61-66 Schimel DS, Parton WJ (1986) Microclimatic controls of nitrogen mineralization and nitrification in sbortgrass steppe soils. Plant and Soil 93:347-357 Schnt~rer J, Clarholm M, Rosswall T (1985) Microbial biomass and activity in an agricultural soil with different organic matter contents. Soil Biol Biochem 17:611-618 Shen SM, Pruden G, Jenkinson DS (1984) Mineralization and immobilization of nitrogen in fumigated soil and the measurement of microbial biomass nitrogen. Soil Biol Biochem 16:437-444 Stevenson FJ (1986) Cycles in Soil: Nitrogen, phosphorus, sulfur and micronutrients, Wiley, New York Trought MCT, Drew MC (1980) The development of waterlogging damage in wheat seedlings (Triticum aestivum L.): II. Accumulation and redistribution of nutrients by the shoot. Plant and Soil 56:187-199 van Veen JA, Ladd JN, Frissel MJ (1984) Modelling C and N turnover through the microbial biomass in soil. Plant and Soil 76:257 - 274 van Veen JA, Ladd JN, Amato M (1985) Turnover of carbon and nitrogen through the microbial biomass in a sandy loam and a clay soil incubated with [14C(U)] glucose and [15N] (NH4)2SO 4 under different moisture regimes. Soil Biol Biochem 17:747-756
Received February 24, 1988