Effects of long-term fertiliser inputs on the quantities of organic carbon in a soil profile under irrigated grazed pasture

Abstract

The long-term fertiliser field trial at Winchmore was used to investigate relationships between pasture production and soil organic carbon (C) storage. In 2009, soil samples to 1 m depth were taken from plots that had been subject to three levels of annual superphosphate fertiliser input for 57 years (nil, 188 kg ha⁻¹ [188PA] and 376 kg ha⁻¹ [376PA]). Although annual pasture production was 2.4–2.8 fold higher for the fertiliser treatments compared with nil P, concentrations and amounts of organic C were similar for the various treatments at most soil depths. Furthermore, differences in total quantities of organic C in the soil profile between the nil P (107 t ha⁻¹), 188PA (101 t ha⁻¹) and 376PA (114 t ha⁻¹) treatments were not significant. The absence of any significant accumulation of soil organic C in response to increased production was attributed to accelerated decomposition of organic matter inputs linked to a combination of improved pasture quality and increased earthworm activity.

Keywords:

• Winchmore
• superphosphate
• grazed pasture
• soil organic carbon
• carbon storage

Introduction

Soil organic matter constitutes the largest terrestrial pool of carbon (C), and predicted climate change is likely to have a major impact on the quantities and dynamics of C in soil (Rodeghiero et al. 2009; Schmidt et al. 2011). Almost all of the organic C added to soil is derived from plant photosynthesis, and is ultimately returned to the atmosphere as carbon dioxide as a result of microbial respiration (Wolf and Wagner 2005; Condron et al. 2010). While it has been shown that amounts of organic C in soil can be substantially altered by shifts in landuse (e.g. Davis & Condron 2002), the relationship between soil C storage and plant production under the same landuse is not well understood. A replicated long-term field trial was established at Winchmore in 1952 to determine fertiliser requirements of irrigated grazed pasture (Condron and Goh 1989; Nguyen et al. 1989). This trial provides an ideal template to test the hypothesis that sustained increases in pasture production result in increased C storage in soil. The specific objective of this study was to determine the cumulative effects of different levels of superphosphate (SSP) fertiliser inputs on quantities of organic C in a soil profile to 1 m depth under irrigated grazed pasture.

Materials and methods

The Winchmore Irrigation Research Station is located in mid-Canterbury, New Zealand (171°48′E, 43°47′S), and was established in 1947 to investigate the management of flood irrigation (border dyke). The soil at Winchmore is a shallow free-draining Lismore stony silt loam (Orthic Brown [New Zealand]; Udic Ustochrept [USDA]), formed from moderately weathered greywacke loess over gravels. The fertiliser trial was initiated in 1952 based on separately fenced and irrigated 0.09 ha plots (borders), and included four replicates of five treatments arranged in randomised blocks. Each treatment was grazed by a separate flock of sheep to avoid nutrient transfer between treatments, and the stocking rates were designed and adjusted to optimise pasture utilisation.

The trial received rainfall (740 mm) plus an average of 4.3 irrigation events (100 mm) per annum (total c. 1150 mm y⁻¹); lime was applied at establishment (5 t ha⁻¹) and in 1972 (4.4 t ha⁻¹) to maintain pH above 6. The treatments selected for this study were the control (nil P), 188 kg SSP ha⁻¹ y⁻¹ (188PA) and 376 kg SSP ha⁻¹ y⁻¹ (376PA). Soil sampling was carried out in April 2009 when the trial had been running for 57 years. Sampling sites were selected at the same relative locations within each of the 12 borders. Due to the nature of the stony soil, sampling using a soil corer was not practicable. Sampling was thus carried out to 1 m depth using a mechanical backhoe to excavate to a depth of 1.5 m. Each pit was approximately 1 m wide by 2 m long. The exposed vertical soil profile was horizontally levelled, and soil and stones were taken from six depths (0–7.5, 7.5–15, 15–25, 25–50, 50–75 and 75–100 cm) using a 40×40×25 cm (0.04 m³) steel frame.

The soil and stones from each depth were weighed and then separated using a combination of sieves from 10 cm to 2 mm. Sub-samples were taken from each depth for analysis, and the residual soil and stones were returned to the pits before refilling. Total C (which equates to organic C in this soil) analysis of finely ground (<150 µm) samples of air-dried soil (20 °C) was carried out using combustion-mass spectrometry (PDZ Europa 20-20 IRMS); the dry soil weight for each depth increment was combined with the organic C concentration to calculate the total quantity of organic C (t C ha⁻¹) in each depth increment and in the soil profile to 1  m. Differences in soil properties between treatments and depths were determined using one-way analysis of variance (ANOVA) with least significant differences (LSD) carried out using Genstat v.11.

Results and discussion

Average weights of soil (oven-dry equivalent) and stones determined for each depth across the 12 plots on the fertiliser trial are presented in Table 1. As expected, the quantity of stones increased with depth and the ratio of soil to stones decreased from 7.7 at 7.5–15 cm to 0.4 at 50–100 cm. The presence of significant stone content below 7.5 cm justified the sampling approach used in this study.

Table 1  Average quantities (t ha⁻¹) of soil and stones present in different depths to 1 m on the Winchmore fertiliser trial plots (n=12;±standard error of mean).

Depth (cm)	Soil (t ha^−1)	Stones (t ha^−1)
0–7.5	692±1.0	0
7.5–15	637±3.6	83±1.8
15–25	900±5.1	180±2.4
25–50	1791±14.8	1416±18.2
50–75	970±4.7	2661±13.3
75–100	1040±10.5	2604±24.6

Soil organic C concentrations (%C) and quantities (t C ha⁻¹) determined in soil taken from the three fertiliser treatments are shown in Table 2. As expected, concentrations and amounts of organic C decreased markedly with soil depth for all treatments from averages of 4.24% and 29.35 t ha⁻¹ at 0–7.5 cm to 0.75% and 7.09 t ha⁻¹ at 75–100 cm respectively. Carbon concentrations were very similar between treatments with no significant differences observed at any depth. The corresponding data for amounts of organic C were more variable but showed a similar trend to organic C concentrations. Nonetheless, the amount of carbon in the topsoil (0–7.5 cm) was significantly higher for the 188PA treatment (30.84 t) compared with nil P (27.96 t), although the difference between 188PA and 376PA (29.25 t) was not significant. The corresponding data for cumulative soil organic C to 1 m confirmed that differences between the nil P (107.07 t), 188PA (100.90 t) and 376PA (114.23 t) treatments were not significant.

Table 2  Mean concentrations (%) and quantities (t ha⁻¹) of organic C determined in soils taken at different depths to 1 m from the nil P, 188PA and 376PA treatments of the Winchmore fertiliser trial after 57 years (2009), including cumulative organic C to 1 m (n=4;±standard error of mean; LSD_0.05=least significant difference at 5% probability).

	Organic C (%)				Organic C (t ha^−1)
Depth (cm)	Nil P	188PA	376PA	P (LSD0.05)	Nil P	188PA	376PA	P (LSD0.05)
0–7.5	4.12±0.025	4.35±0.087	4.25±0.150	0.333 (0.143)	27.96±0.567	30.84±0.725	29.25±0.579	0.030 (2.008)
7.5–15	3.10±0.061	3.01±0.099	2.91±0.090	0.346 (0.271)	20.60±3.789	17.77±1.486	17.91±2.123	0.709 (8.480)
15–25	2.14±0.114	2.04±0.064	2.22±0.105	0.478 (0.309)	20.63±1.893	17.26±1.430	20.13±3.337	0.574 (7.560)
25–50	1.35±0.181	1.22±0.148	1.38±0.200	0.796 (0.568)	21.33±3.965	19.06±2.212	28.53±6.178	0.332 (14.160)
50–75	1.10±0.059	1.07±0.122	1.05±0.069	0.922 (0.281)	9.07±0.802	10.49±0.544	10.09±1.260	0.553 (2.937)
75–100	0.76±0.080	0.76±0.092	0.73±0.052	0.962 (0.245)	7.48±0.760	5.48±0.850	8.32±2.287	0.411 (4.719)
Cumulative to 1 m					101.70±10.264	100.90±3.899	114.23±12.662	0.636 (30.960)

Average annual dry matter yields measured between 1954 and 1985 were 4.0, 9.82 and 11.1 t ha⁻¹ for the nil P, 188PA and 376PA treatments respectively (Nguyen et al. 1989). This 2.4–2.8 fold increase in pasture production is likely to have increased inputs of organic C to soil in the form of herbage and root residues and animal faeces (and urine) under the fertilised treatments compared with nil P (Kemp et al. 2000). However, the findings of this study clearly demonstrated that, despite increased C inputs over 57 years, levels of soil organic C in under-fertilised pasture were similar to non-P-fertilised treatments. This suggests that accelerated decomposition of organic matter inputs occurred in fertilised soils compared with the nil P treatment soil, which may in turn be related to a combination of factors including improved pasture quality and increased earthworm activity. Thus, Nguyen et al. (1989) reported that the clover content of pasture was markedly higher in the 188PA and 376PA treatments (26%) compared with nil P (13%), while the proportion of low-quality weed species was higher in the nil P treatment (23%) than in the fertiliser treatments (10%). Furthermore, Fraser et al. (1994) found that numbers and biomass of earthworms were significantly higher in the 188PA and 376PA treatment soils compared with the nil P soil.

It has also been shown that the soil microbial biomass (measured as quantity of DNA per g of soil) is similar in the 0–7.5 cm soil depth across all treatments at the site (P=0.784) (SA Wakelin, unpublished data). As such, despite increased inputs of photosynthetic C into the soil, a constant soil microbial biomass is retained. Given equivalent microbial biomass, rates of soil C mineralisation in the fertilised plots must have been greater than in the unfertilised soils. Thus, either the size of the microbial communities was not a constraint on C mineralisation, or the availably of other nutrients (such as P) or other factors increases the net rate of C cycling in the fertilised soils (altered metabolic status). These hypotheses are being tested in ongoing experiments.

Conclusions

The findings of this study demonstrated that while fertiliser inputs increase pasture production, concomitant improvements in pasture quality and soil biological activity lead to enhanced decomposition of organic inputs and thus limited accumulation of organic C in the soil profile.

Acknowledgements

We thank past and present staff at Winchmore for initiating and maintaining the long-term fertiliser trial. In particular, we thank John Carson and Dr Todd White for their assistance with facilitating and organising the soil sampling in 2009. Special thanks are also due to Robyn Damary-Homan and Jemma Mackenzie for their outstanding efforts in soil sampling at Winchmore. Funding for this study was provided by the Agricultural and Marketing Research and Development Trust (AGMARDT), the New Zealand Ministry of Agriculture and Forestry, and Environment Canterbury.