Soil organic carbon, nitrogen, and phosphorus 13 yr after abruptly disturbing Northern Great Plains grassland

Abstract: Thirteen years after cultivating native grassland and establishing continuous wheat (Triticum aestivum L.) and wheat–fallow rotations in southern Alberta, surface soil total N levels were 15% lower, and nitrate (60–90 cm) concentrations were 2.5- and 17-fold greater, than native grassland. Wheat–fallow, even without fertilization, markedly enhanced potential nitrate loss through the root zone.

The first results of a study designed to assess the effect of abruptly altering the steady-state condition of native grassland on soil properties were reported by Dormaar and Willms (2000). They found that 2 yr after cultivating this grassland, soil organic C, N, and P concentrations were all significantly reduced. Five years after the grassland was cultivated, surface soil total C and N concentrations under the seeded perennial grasses were 30% and 25% lower than the undisturbed native grassland soil, respectively (Whalen et al. 2003). Thirteen years after cultivation, perennial grasses restored surface soil organic C storage to levels of the native grassland, whereas continuous wheat (Triticum aestivum L.) and wheat-fallow remained 20% lower (Wang et al. 2010). Although C has been the focal point of most research, uncertainty remains about how abrupt land-use change alters total N, P, and the depth distribution of soil nitrate  and soil test P after decadal or more timescales in the Northern Great Plains of southern Alberta. Here, we present soil organic C, total N and P, nitrate, and soil test P concentrations 13 yr after the initial cultivation in response to abandonment, perennial grass establishment, and continuous wheat or wheat-fallow rotation to a soil depth of 90 cm.
Treatment plots (3 m × 10 m) were established on the previously uncultivated native grassland in spring 1994 by cultivating with a rototiller. The treatments were arranged in a randomized complete block design (RCBD) with four blocks. The seven treatments were (1) cultivated and then abandoned, leaving the plot to reestablish through succession from residual germplasm with no harvesting (abandoned); (2) cultivated and seeded with perennial monoculture of crested wheatgrass (Agropyron cristatum L. Gaertn.); (3) cultivated and seeded with perennial monoculture of Russian wildrye (Psathyrostachys juncea Fisch.); (4) annually cultivated continuous spring wheat (Triticum aestivum L.); (5) biannually cultivated spring wheat-summerfallow rotation (wheat-fallow); (6) a native grassland that was annually harvested (harvested grassland); and (7) a native grassland that was left undisturbed as a control (native grassland). The crested wheatgrass, Russian wild rye, and spring wheat were seeded at 15 cm row spacing. The experimental site was fenced to exclude cattle, deer, and small herbivores. A 5 m buffer was established between the fence and plots. In early May, the spring wheat treatments were cultivated using a Triple-K cultivator and packed with sheep's foot packers, annually for continuous wheat and during the wheat year for wheat-fallow. The plots were seeded (78.4 kg seed ha −1 ) with a three-point pony drill. In-crop weeds were controlled with selective herbicides and the wheat-fallow rotation was treated once with herbicide per fallow season (chemical fallow). Fertilizer and supplemental water were not applied. All treatments were annually harvested (except for the native grassland and the abandoned treatmentsnot harvestedand the wheatfallow treatmentharvested only during the wheat year) with a forage harvester at a height of 8 cm at peak standing crop in early July or August. Crested wheatgrass was harvested about 1 mo earlier because of its earlier phenological development.
Soil samples were randomly collected from each plot using a 6.7 cm diameter soil probe in early July 2006, which was the fallow year for the wheat-fallow rotation. Four samples were collected to a depth of 90 cm and divided into 0-15, 15-30, 30-60, and 60-90 cm soil layers. Coarse roots (>2 mm) were removed, and soil samples were composited for each soil layer. Soil samples were then ground to pass through a 2 mm sieve to measure soil NO 3 -N and soil test P determination, whereas subsamples were ground to pass through a 0.15 mm sieve to measure total C, inorganic C, total N, and P. Soil NO 3 -N was extracted with 2 mol L −1 KCl (1:5 soil to KCl) and its concentration determined with an automated colorimeter (AutoAnalyzer III, Bran + Luebbe, Norderstedt, Germany). Soil test P was determined using the Kelowna extraction method with a 1:5 soil to extract ratio (Van Lierop 1988). The soil total C and N concentrations were determined by dry combustion (Carlo Erba, Milan, Italy). Inorganic C was determined using a modified method of Amundson et al. (1998), and soil organic C was calculated as the difference between soil total C and inorganic C. Soil total P concentrations were determined by digestion with 18 mol L −1 H 2 SO 4 (Parkinson and Allen 1975). The P concentration in the total P digest and the Kelowna extract (soil test P) were determined with an autoanalyzer (Astoria, Clackamas, OR, USA).
Statistical analyses were conducted with SAS 9.1 software (SAS Institute Inc. 2005). The UNIVARIATE procedure was used to check the residuals for normality and identify outliers. When an outlier was detected, it was removed before the ANOVA was performed. All data were analyzed by one-way ANOVA as a RCBD with the MIXED procedure. Depth was a repeated measure. Means were compared by Fisher's protected LSD test at α = 0.05.
Soil organic C concentrations were significantly affected by the land-use change (P = 0.004), soil depth (P < 0.001), and land-use change × sampling depth (P < 0.001; Table 1). For the surface soil (0-15 cm), converting native grassland to continuous wheat or wheatfallow reduced soil organic C by 23% compared with the undisturbed native grassland. Cultivating native grassland followed by abandonment reduced soil organic C by 11% compared with the native grassland, but abandonment led to 13% greater soil organic C than both wheat cropping systems. In contrast, even without returning the aboveground residues, converting native grassland to perennial grasses or harvesting the native grassland did not significantly alter soil organic C compared with the native grassland 13 yr after cultivation. This indicates that the belowground residues were mostly responsible for maintaining the soil organic C pool under the two perennial grasses and the harvested native grassland. For the 15-30 cm soil layer, the organic C concentration was 13% lower with continuous wheat than the native grassland, but no other differences were Note: Numbers in a row followed by different lowercase letters indicate significant land-use change effect, and numbers in a column followed by different uppercase letters indicate significant soil depth effect (P < 0.05). significant among land-use changes at the 15-30, 30-60, or 60-90 cm depths.
The soil total N concentrations were significantly affected by the sampling depth (P < 0.001) and sampling depth × land-use change (P < 0.001) but not land-use change alone (P = 0.069; Table 1). In the 0-15 cm soil layer, the cultivated then abandoned soil had 8% lower total N than the native grassland, whereas the continuous wheat and wheat-fallow led to 7%-15% lower soil total N than all other treatments (Table 1). In the 15-30 cm soil layer, wheat-fallow led to 13% greater soil total N than continuous wheat. In the 30-60 cm soil layer, no significant differences among treatments were detected, whereas the wheat-fallow rotation had 13% less total N than the crested wheatgrass in the 60-90 cm soil layer, but no other differences were detected among treatments.
The soil total P concentrations were not significantly affected by land-use change (P = 0.913) or land-use change × soil depth (P = 0.613) but were significantly affected by depth (P < 0.001). The soil total P concentration was lowest in the 15-30 cm soil layer (0.32 g P kg −1 ), highest in the 30-60 and 60-90 cm soil layers (0.40 and 0.42 g P kg −1 , respectively), and in between these values in the 0-15 cm soil layer (0.35 g P kg −1 ).
The soil NO 3 -N and soil test P concentrations were significantly affected by land-use change, sampling depth, and land-use change × sampling depth (All P < 0.003; Fig. 1). The soil NO 3 -N concentrations were disproportionately greater in the 30-60 and 60-90 cm soil layers with continuous wheat and wheat-fallow than the other treatments. The soil NO 3 -N concentration with continuous wheat was 6.3-fold greater at 30-60 cm and 2.5-fold greater at 60-90 cm than the native grassland, and the wheatfallow was 8.5-fold greater at 30-60 cm and 17-fold greater at 60-90 cm than the native grassland. Wheat-fallow led to 4.1-fold greater soil NO 3 -N concentration at 60-90 cm than continuous wheat. The surface-soil test P concentrations were greatest with wheat-fallow followed by the continuous wheat. There were minimal soil test P differences among treatments within deeper soil layers.
Although the soil organic C was initially depleted 5 yr after cultivation with every land-use change relative to the native grassland (Whalen et al. 2003), after 13 yr, the soil organic C concentrations were restored to the native grassland levels in most treatments except the continuous wheat and wheat-fallow rotation. As expected, there was higher soil organic C and total N concentrations in the surface soil, reflecting greater plant biomass input near the soil surface through roots Fig. 1. Soil NO 3 -N (a) and soil test P (Kelowna-extractable P) (b) concentrations 13 yr after abrupt grassland-use change in the Northern Great Plains of southern Alberta. For each soil layer, bars with different lowercase letters are significantly different (P < 0.05). Error bars indicate ± standard error. Note: Bars follow legend order. and aboveground biomass remaining after harvest (<8 cm standing crop residue).
The soil organic C and total N concentrations in the surface soil for the abandoned treatment were 89% and 92% of those for the native grassland, respectively. Thus, 13 yr was insufficient to completely recover from the one-time intensive cultivation and removal of the aboveground biomass. Compared with the native grassland, 13 yr of continuous wheat or wheat-fallow cropping resulted in 23% less soil organic C and 15% less total N, but no significant difference in total P concentrations in the 0-15 cm soil layer, clearly demonstrating that continuous wheat or wheat-fallow affected soil organic C and total N concentrations more than total P. Five years after cultivation, Whalen et al. (2003) reported that for continuous wheat and wheat-fallow, soil total C was 38% and 39% lower, respectively, and soil total N was 30% and 31% lower, respectively, than total C and total N of native grassland soil. Comparing the 5 and 13 yr data, the soil organic C and total N gaps between the soils cropped to wheat and the native grassland were reduced by about 16 and 15 percentage points. This provides evidence that some of the soil organic C and N lost with the initial disturbance was replenished with continuous wheat and wheat-fallow at similar rates with annual and biannual cultivations. This is inconsistent with studies that found reduced fallow increased the rate C was sequestered in southern Alberta (Bremer et al. 2002;Smith et al. 2016), based on the similar apparent rate soil organic C and total N concentrations increased with both continuous wheat and wheat-fallow in the 0-15 cm soil layer. This discrepancy may be explained by the initial breaking of the soil (Campbell et al. 1975;Woods et al. 2013) having a more significant impact on the recovery of soil organic C and total N than the annual or biannual cultivation of the continuous wheat or wheat-fallow over the 13 yr period. This implies that more time may be required to observe differences in soil organic C and total N concentrations between continuous wheat and wheat-fallow after the initial breaking of the soil. Similar soil total P concentrations with continuous wheat and the native grassland are comparable to the results reported by Whalen et al. (2003). However, 5 yr after cultivation, the wheat-fallow and the abandoned soil had 8% and 9% lower total P than the native undisturbed grassland, respectively (Whalen et al. 2003). After 13 yr, soil total P differences were no longer evident. This implies that the initial cultivation of this grassland reduced soil total P, and the wheat-fallow and abandoned treatments restored soil total P levels between 5 and 13 yr after initial cultivation.
A slightly but significantly greater total N concentration in the 15-30 cm soil layer with wheat-fallow than continuous wheat was probably due to biannual cultivation of wheat-fallow instead of annually for continuous wheat. Biannual cultivation likely slowed mineralization of organic matter, leaving more total N (mostly organic N) in the fallow treatment (Six et al. 1998). The significant accumulation of soil NO 3 -N outside the apparent root zone (30-60 and 60-90 cm soil layers) with continuous wheat and wheat-fallow indicates that converting grassland to these cropping systems increases the risk for plant-available N loss compared with the perennial grasses and the native grassland. In particular, the starkly greater soil NO 3 -N concentration at 60-90 cm depth with wheat-fallow than continuous wheat clearly shows that wheat-fallow retained significantly less soil NO 3 -N in the effective root zone (Fig. 1); even without fertilization. For a frame of reference, this is roughly the equivalent of 44 kg NO 3 -N ha −1 (measured bulk density of 1.42 g cm −3 ); the ultimate fate is likely groundwater.
In conclusion, introducing perennial monoculture grassland restored soil organic C and total N to levels of the native grassland 13 yr after cultivation and seeding. However, converting native grassland to a wheat cropping system significantly reduced soil organic C and total N concentrations. Cultivating native grassland followed by abandonment also led to lower soil organic C and total N concentrations than undisturbed native grassland, but abandonment led to smaller C and N losses than continuous wheat and the wheat-fallow rotation. The soil total P levels were restored as indicated by similar values for all land uses, and this indicates that P is more resilient to land-use change than C and N. Continuous wheat and wheat-fallow had large proportions of soil NO 3 -N in the 60-90 cm soil layer, indicating that even without fertilization these land-use changes increased potential N loss through the root zone.