Pasture renewal on Bay of Plenty and Waikato dairy farms: impacts on pasture performance post-establishment

Abstract

To determine the impact of pasture renewal on dairy pasture performance, a total of 24 renewed and unrenewed pastures were monitored in Bay of Plenty and Waikato for 5 years. Renewed pastures produced an additional 1.5, 1.8 and 1.9 t dry matter (DM) ha⁻¹ in the first, second and third years of monitoring, respectively, compared with unrenewed pastures. There was a greater contribution of clover, sown grasses and unsown grasses and a smaller contribution of broadleaf weeds in renewed than unrenewed pastures in some years (% of total DM). The sown grass DM content (kg DM ha⁻¹), perennial ryegrass tiller densities and endophyte infection frequencies were similar in renewed and unrenewed pastures. The abundance of invertebrate pests and total nematodes was lower in renewed than unrenewed pastures in some years. The greater clover content, fewer weeds and fewer insect pests, contributed to the greater herbage production of renewed pastures.

Keywords:

• dairy pastures
• DM production
• pasture persistence
• pasture quality
• pasture renewal

Introduction

Pasture renewal can improve pasture performance, livestock productivity and farm profitability on dairy farms (Brazendale et al. 2011; Shalloo et al. 2011). Some of these benefits are derived from improved plant genotypes but the reduction in weed and pest burdens is also important. During the renewal process, weed populations can be controlled and insect pest population cycles can be disrupted through cultivation and physical soil disruption as well as by removing host plants such as Poa spp. or summer grasses that provide feed for pests such as black beetle (Heteronychus arator [F.]) and Argentine stem weevil (Listronotus bonariensis [Kuschel]) (Bell et al. 2006; Zydenbos et al. 2011).

In New Zealand, dairy farmers have expressed concern that pasture renewal benefits are not being realised on-farm and that pastures are rapidly reverting to unsown species (Bewsell et al. 2008; Kelly et al. 2011; Tozer et al. 2011). This is particularly the case in the upper North Island, where frequent droughts have occurred and where pastures have been subjected to severe insect pest and weed pressures (e.g. Tozer et al. 2012; Gerard et al. 2013; Porteous & Mullan 2013). In addition, farm systems have become more intensive, with higher stocking rates and larger livestock, placing further biotic stresses on pastures (Macdonald et al. 2011). Farmers are also concerned that yields gained from modern cultivars grown in commercial plot trials do not correlate well with yields gained on-farm. This has led to farmer scepticism regarding the purported benefits of pasture renewal and concerns regarding pasture persistence (Bewsell et al. 2008). In a survey of 800 Waikato and Bay of Plenty (BoP) dairy farmers, pasture persistence was given as one of the main factors limiting pasture performance and farmers became less satisfied with their pastures as they aged (from 1–3 years after sowing [Brazendale et al. 2011]).

To determine the impact of renewal on pasture performance, a project was established on commercial dairy farms in Waikato and BoP. Herbage production of renewed and unrenewed pastures was monitored over 5 years in conjunction with factors associated with persistence, including tiller densities, weed and invertebrate populations and endophyte infection frequencies. We hypothesised that herbage production would be greater in renewed than unrenewed pastures and that herbage production benefits would remain for 5 years after sowing.

Methods

Design

Two major dairy regions were chosen for this on-farm comparison of pasture renewal effects: Waikato and BoP, which together contained 28% of New Zealand’s dairy cattle and produced 27% of New Zealand’s milk solids in 2012–13 (DairyNZ 2013). Five commercial farms were selected in each region (Fig. 1). The farms were located between Tokoroa and Gordonton in Waikato and on the Edgecumbe plains in BoP. On each farm, two or three paddocks were chosen that resulted in seven renewed pastures and five unrenewed pastures in each of the two regions, giving a total of 24 pastures. Criteria for selection within a farm included similarity in soil type and previous management histories. Pastures selected were considered by the manager to be typical of older pastures on their farms. The paddock for renewal was randomly selected by the farmer, except on two Waikato farms where the lowest performing paddocks were chosen for renewal, and these were compared with an ‘average paddock’ that remained unrenewed. Across regions, this comparison provides a contrast between a new pasture and an ‘average old’ pasture. Once established, renewed and unrenewed pastures were managed similarly (including decisions regarding fertiliser application and grazing management). Approximately 180 kg N ha⁻¹ was applied annually on each farm during the study period, with the timing and number of split-applications each year varying from farm to farm (S. Beynon, BoP farm consultant, unpubl. data).

Figure 1 Location of the five farms in each of Waikato and Bay of Plenty where a total of 24 renewed and unrenewed pastures were monitored.

Pasture renewal strategies

Pastures were renewed in 2007, 2008, 2009 or 2010 (Table 1). On nine farms, pastures were renewed with diploid or tetraploid perennial ryegrass (Lolium perenne L.) containing a novel endophyte (AR37, AR1 or NEA2) and white clover (Trifolium repens L.) with proprietary seed treatments recommended by the seed suppliers. Some of the pastures went through a cropping phase (maize (Zea mays L.), turnips (Brassica rapa L.) or triticale (× Triticosecale Wittm. ex A. Camus) that included soil cultivation. The unrenewed pastures were greater than 10 years old with unknown cultivars containing standard endophyte, except for one comparison with a 5 year old paddock containing AR1 endophyte. On one farm in Waikato, renewed paddocks sown with tall fescue (Schedonorus arundinaceus [Schreb.] Dumort., nom. cons.) containing MaxP endophyte were compared with unrenewed pastures of older endophyte-free tall fescue. All pastures were sown using a direct drill.

Table 1 Year of renewal, renewal sequence (crop-to-grass or grass-to-grass) and cultivar-endophyte combination sown in 14 renewed pastures monitored on five farms in each of Waikato and Bay of Plenty (BoP). The crops sown were maize (Zea mays L.), turnips (Brassica rapa L.) or triticale (× Triticosecale Wittm. ex A.Camus).

Region	Farm	Year of renewal	Renewal sequence	Cultivar sown
Waikato	Farm 1	2009	Grass-to-grass	Extreme AR37
	Farm 2	2007	Grass-to-grass	Arrow, AR1
		2008	Grass-to-grass	Bealey/Arrow, NEA2/AR1
	Farm 3	2009	Crop (triticale)	Arrow/Alto AR1
	Farm 4	2010	Crop (maize)	Alto AR37
	Farm 5	2007	Crop (turnips)	Tall fescue Max P
		2009	Crop (turnips)	Tall fescue Max P
BoP	Farm 1	2008	Grass-to-grass	Bealey NEA2
		2009	Grass-to-grass	Bealey NEA2
	Farm 2	2008	Crop (maize)	Bealey NEA2
		2009	Grass-to-grass^a	Bealey NEA2 irrigated
	Farm 3	2008	Grass-to-grass^a	Commando AR37
	Farm 4	2008	Grass-to-grass	Extreme AR37
	Farm 5	2008	Crop (maize)	Bealey NEA2

^aPaddocks were cultivated before sowing the pasture.

Herbage accumulation and botanical composition

Herbage accumulation and botanical composition were assessed from August 2009 until August 2014. Three grazing exclusion cages were randomly positioned in each pasture and moved to a new randomly selected position after above ground vegetation was harvested within a 0.125 m² quadrat to ‘grazing height’ (c. 2 cm) from within each cage, using an electric shearing hand-piece. Harvested vegetation was bulked, mixed thoroughly and oven dried at 65 °C until a constant weight was reached (c. 48 h) and the dry weight was used to calculate total dry matter (DM) (kg DM ha⁻¹). Harvests occurred at intervals of between 4 and 8 weeks in all pastures depending on the season. Seasons were defined as follows: summer, December to February; autumn, March to May; winter, June to August; and spring, September to November.

To determine the botanical composition of the pasture, a subsample of approximately 100 g of the freshly harvested vegetation was sorted into sown and unsown grasses, clover species (Trifolium spp.) and broadleaf weeds and each component was oven dried separately. Each component was weighed to estimate its contribution to total DM (expressed as a percentage) and herbage accumulation.

Tiller densities

Assessments were undertaken biannually in autumn (late April/May) and mid-spring (late September/October) from spring 2011 until autumn 2014. Twenty soil cores of 10 cm in diameter were removed at regular intervals along a transect adjacent to the pasture cages. Perennial ryegrass and unsown grass tillers were counted in each core to estimate tiller density (m⁻²). Tiller densities of pastures sown with tall fescue were not included in the statistical analyses.

Endophyte presence

Assessments were undertaken each November from 2010 until 2013. A healthy vegetative adult tiller from each of 100 randomly selected perennial ryegrass plants sampled along a diagonal transect that extended the length of each paddock was cut to ground level. Frequency of sampling (i.e. distance between sampled plants) depended on the length of the transect and was proportional to the length of the paddock. Collected tillers were taken to the laboratory and refrigerated until being recut and pressed on to 0.45 μm nitrocellulose blotting paper (Protran BA45, Thermo Fisher Scientific) for 2–3 sec within 24 h of sampling. Blotting sheets were processed using the tissue print immunoassay method of Hahn et al. (2003) to determine the frequency of infection by an endophyte.

Invertebrate species presence and abundance

Invertebrate assessments were made each February from 2009 until 2014. A modified garden blower-vacuum with a 10 cm diameter collection sleeve was used to collect above-ground invertebrates along a randomly selected 30 m transect in the vicinity of the pasture growth cages. Collected invertebrates were kept in cool storage until hand-sorting in the laboratory. All invertebrates were counted and identified to the lowest taxonomic rank possible.

To assess below-ground invertebrate abundance, 10 spade squares (20 × 20 cm to 20 cm depth) were removed at 2 m intervals along the 30 m transect. Soil from each spade square was hand crumbled and all soil invertebrates visible to the naked eye were removed, identified and counted in the field.

Nematodes

Assessments were undertaken each February from 2009 until 2014. A soil core (2.5 cm diameter, 10 cm depth) was sampled from beside each spade square that had been removed for the soil invertebrate sampling. The 10 cores were hand crumbled and bulked for each paddock and a 100 g subsample of soil was removed. Nematodes were extracted using the method described by Bell & Watson (2001), counted and plant parasitic forms identified to genus level under a stereo microscope at 40–80× magnification. Soil moisture was determined gravimetrically (Evett 2008) from the remaining portion of the hand-crumbled soil and results used to calculate the number of total, root knot (Meloidogyne spp.), cyst (Heterodera spp.), lesion (Pratylenchus spp.) and pin (Paratylenchus spp.) nematodes g⁻¹dry soil.

Soil nutrients

Assessments were undertaken each spring (September or October) from 2011 until 2013. Twenty-five cores (2.5 cm diameter × 7.5 cm depth) were removed every 1.0–1.5 m along a randomly selected 30 m transect. Cores were bulked for each paddock and sent to a commercial laboratory for soil analyses (pH, Olsen P, K, S[SO₄], Mg, Ca and Na).

Climate data

Long-term rainfall and temperature data were obtained from weather stations central to the farms monitored in Waikato (Arapuni) and BoP (Edgecumbe). Air temperatures, soil temperatures, total solar radiation and rainfall data for the experimental period were obtained for each of the farm locations based on the National Institute of Water and Atmospheric Research virtual climate database (Tait et al. 2006; Tait & Woods 2007), using smoothed interpolated actual climate data.

Statistical analyses

Data were analysed using residual maximum likelihood (REML) (Patterson & Thompson 1971) in GenStat, 13th edition (GenStat 2010). REML was used since it was likely that differences between paddocks within a farm were less than differences between farms and because the design was unbalanced due to the varying number of paddocks per farm. The REML fixed effects were region, renewed or unrenewed and their interaction. The REML random effects were farm and paddock within farm. Analyses were undertaken for each individual year and also for the total period. Since the number of paddocks sampled varied between farms, the predicted means for seasons do not sum exactly to the predicted total means (Table 2). The nematode and soil insect data were log-transformed to normalise the variance so that the normality assumptions of the analyses were met. All reported P values are calculated from transformed data but arithmetic means are presented.

Table 2 Seasonal and total annual dry matter production (kg DM ha⁻¹) in renewed and unrenewed Waikato and Bay of Plenty (BoP) dairy pastures from September 2009 until February 2014. Renewed pastures were 1 or 2 years old when measurements began.

	Renewed				Region
Season	Yes	No	SED	P value	Waikato	BoP	SED	P value
2009–2010
Spring	4390	4280	274	0.701	4260	4410	431	0.686
Summer	6330	5680	334	0.074	5800	6210	832	0.644
Autumn	2820	2730	258	0.754	2120	3430	675	0.093
Winter	2750	2060	112	<0.001	2140	2670	372	0.152
Annual	16220	14750	388	0.003	14250	16710	2061	0.256
2010–2011
Spring	5860	5340	347	0.159	4830	6380	897	0.106
Summer	6530	5440	349	0.008	5050	6910	544	0.006
Autumn	3160	2950	299	0.494	3180	2930	577	0.695
Winter	2450	2390	237	0.817	2160	2670	362	0.172
Annual	17960	16120	604	0.010	15190	18880	1893	0.071
2011–2012
Spring	6230	5460	289	0.020	6040	5650	433	0.326
Summer	4760	4350	255	0.134	3660	5450	779	0.051
Autumn	4240	3830	342	0.258	4080	3990	697	0.870
Winter	2710	2340	202	0.086	2290	2760	409	0.270
Annual	17870	15980	739	0.025	16030	17820	2064	0.427
2012–2013
Spring	4620	4610	347	0.985	4130	5110	422	0.065
Summer	4590	3820	433	0.101	3210	5200	671	0.015
Autumn	3380	2940	431	0.323	2960	3360	502	0.530
Winter	2200	1980	168	0.207	1470	2720	324	0.006
Annual	14760	13350	707	0.069	11710	16410	1363	0.010
2013–2014
Spring	4860	5320	260	0.098	4410	5770	420	0.014
Summer	5210	5100	345	0.761	4450	5850	588	0.042
Autumn	3350	3350	200	0.977	1750	4950	611	<0.001
Winter	2690	2380	243	0.221	2480	2590	470	0.785
Annual	16090	16150	729	0.939	13070	19170	1663	0.007
2009–2014
Spring	25910	25020	816	0.295	23590	27340	2058	0.109
Summer	27440	24390	1181	0.024	22160	29660	2783	0.025
Autumn	16830	15810	816	0.235	14010	18620	2643	0.129
Winter	12740	11140	288	<0.001	10520	13370	1682	0.119
Total	82830	76350	1954	0.006	70180	88990	8306	0.052

Since the number of pastures varies for each farm, the predicted means for season do not sum exactly to the predicted annual means.

Summer, December–February; autumn, March–May; winter, June–August; spring, September–November.

SED, standard error of difference.

Results

Climate

Waikato and BoP typically have warm, wet summers with a similar long-term rainfall (Waikato: summer = 300 mm, annual = 1396 mm; BoP: summer = 305, annual = 1366 mm) and average annual maximum air temperatures (averaging 19.1 °C across both regions, Fig. 2). The minimum average annual air temperature is lower in Waikato than BoP (8.2 vs 9.1 °C), resulting in a marginally lower average annual temperature in Waikato than BoP (13.6 vs 14.2 °C, Fig. 2).

Figure 2 Average monthly rainfall (■), average monthly temperature (—) and average monthly minimum and maximum temperatures (– – –) for: A, Arapuni (Waikato); B, Edgecumbe (Bay of Plenty), from 1971 until 2000.

Generally, BoP was warmer, wetter and sunnier than Waikato over the assessment period. This corresponded to greater average soil temperatures in BoP than Waikato in summer 2010, 2011 and autumn and spring 2013 (P < 0.05, data not shown). Total solar radiation was greater in BoP than Waikato for all seasons assessed (P < 0.05), except for autumn and winter 2009 and 2012 (data not shown, P > 0.05).

In Waikato, summer rainfall in 2012–2013 (150 mm) and 2013–2014 (158 mm) was much lower than the long-term average (300 mm). In BoP, summer rainfall in 2012–2013 (166 mm) was also much lower than the long-term average (305 mm).

Herbage production on a seasonal and annual basis

When summed over the 5 year study, herbage production in summer was greater in renewed than unrenewed pastures by 3050 kg DM ha⁻¹, in winter by 1600 kg DM ha⁻¹ and overall by 6480 kg DM ha⁻¹ (Table 2). Total annual herbage production was greater in renewed than unrenewed pastures, equating to an additional 1470 kg DM ha⁻¹ in 2009–2010, 1840 kg DM ha⁻¹ in 2010–2011 and 1890 kg DM ha⁻¹ in 2011–2012. Renewed pastures produced an additional 690 kg DM ha⁻¹ in winter 2010, 1090 kg DM ha⁻¹ in summer 2010–2011 and 770 kg DM ha⁻¹ in spring 2011, when compared with unrenewed pastures (Table 2). However, there were no significant differences between renewed and unrenewed pastures in seasonal or total annual herbage production in 2012–2013 or 2013–2014.

Herbage production was greater in BoP than Waikato pastures in summer by 7500 kg DM ha⁻¹ when summed over the 5 years (Table 2). Total annual herbage production was also greater in BoP than Waikato pastures in 2012–2013 by 4700 kg DM ha⁻¹ and 2013–2014 by 6100 kg DM ha⁻¹ (Table 2).

When compared with pastures in Waikato, pastures in BoP produced an additional 1860 kg DM ha⁻¹ in summer 2010–2011, 1990 kg DM ha⁻¹ in summer 2012–2013, 1250 kg DM ha⁻¹ in winter 2013, 1360 kg DM ha⁻¹ in spring 2013 and 1400 kg DM ha⁻¹ in summer 2013–2014 (Table 2).

There were several interactions between season and region. When compared with unrenewed pastures, renewed pastures produced an additional 1140 kg DM ha⁻¹ in winter 2010 in BoP (P = 0.002) but not in Waikato (averaging 2140 kg DM ha⁻¹ across both treatments). Similarly, in 2010–2011, renewed pastures produced an additional 3560 kg DM ha⁻¹ in BoP (P = 0.010) but no difference between renewal treatments were observed in Waikato (averaging 15,190 kg DM ha^–1, P > 0.05).

Sown and unsown species DM content

When DM production data for each of the seasons was summed over the 5 years (2009–2014), sown grasses production was greater in renewed than unrenewed pastures in winter by 1270 kg DM ha⁻¹ (P = 0.005) and of clover in winter by 390 kg DM ha⁻¹ and spring by 870 kg DM ha⁻¹ (P < 0.05), but broadleaf weed production was lower in renewed than unrenewed pastures in winter by 360 kg DM ha⁻¹ (P = 0.002, Fig. 3). Dry matter production of clover summed over all seasons was also greater in renewed than unrenewed pastures by 3030 kg DM ha⁻¹ (P = 0.022). There were no significant effects of renewal on total annual DM production of sown grasses (Fig. 4). Total annual DM production of clover was significantly greater in renewed than unrenewed pastures by 480 kg DM ha⁻¹ in 2010–2011 and for unsown grasses by 700 kg DM ha⁻¹ in 2012–2013, but lower for broadleaf weeds in renewed than unrenewed pastures by 580 kg DM ha⁻¹ in 2010–2011 (Fig. 4).

Figure 3 Seasonal dry matter production (kg DM ha⁻¹) summed over the 5 year period from 2009–2014 of: A, sown grass; B, unsown grass; C, clover; D, broadleaf weed in renewed (■) vs unrenewed (■) pastures. Data are averaged over both regions. Bars are the standard error of difference.

Figure 4 Annual production of: A, sown grass; B, unsown grass; C, clover; D, broadleaf weed in renewed (■) vs unrenewed (■) dairy pastures in Waikato and Bay of Plenty, from September 2009 until August 2014 (kg DM ha⁻¹). Data are averaged over both regions. Bars are the standard error of difference.

When summed over 5 years, broadleaf weed DM production was greater in BoP than Waikato in spring by 1210 kg DM ha^–1 (P = 0.027). Total annual DM production of sown grasses was also greater in BoP than Waikato in 2012–2013 by 3590 kg DM ha⁻¹ (P = 0.040) and in 2013–2014 by 4110 kg DM ha⁻¹ (P = 0.048), as was clover by 1840 kg DM ha⁻¹ (P = 0.031) and broadleaf weeds by 880 kg DM ha⁻¹ (P = 0.038).

There was an interaction for annual DM production of clover in 2009–2010. Total annual dry matter production of clover was greater in renewed than unrenewed pastures in BoP by 1650 kg DM ha⁻¹ (1870 vs 220, P = 0.034) but similar in renewed and unrenewed pastures in Waikato (averaging 1350 kg DM ha⁻¹).

Although most renewed pastures produced more DM than their unrenewed ‘control’ pasture, Fig. 5 demonstrates how the renewal effect (i.e. difference in pasture production between the renewed pasture and the unrenewed ‘control’ pasture on the same farm) varied greatly between farms and years. Some renewed pastures achieved a gain in DM production relative to their control in one year, but sustained a loss relative to their control unrenewed pasture in the following year. Renewed pastures in BoP consistently performed better than those in Waikato with seven occasions where renewed produced less DM than unrenewed pastures, compared with 15 occasions in the Waikato.

Figure 5 Effect of pasture renewal on total annual DM production from 2009 to 2014 (t DM ha⁻¹) in Bay of Plenty (grey bars) and Waikato (hatched bars). Each bar represents the difference in pasture production between a renewed and ‘control’ unrenewed pasture on the same farm. A positive number indicates that the renewed pasture produced more than the unrenewed ‘control’ pasture and a negative number indicates that the renewed produced less than the unrenewed pasture, within any given farm and year. The order of farm pasture comparisons is the same for all years.

Botanical composition

Trends in the contribution of clover and broadleaf weed were consistent with their trends in DM production, with a higher percentage of clover and lower proportion of broadleaf weeds in renewed than unrenewed pastures for at least one of the years assessed (P < 0.05, Fig. 6). There was also an interaction between renewal and region for clover in 2009–2010, with a higher contribution of clover in renewed than unrenewed pastures in BoP (10% vs 1%) but a similar contribution between treatments in Waikato (averaging 11% of total DM, P = 0.014).

Figure 6 Botanical composition of: A, renewed; B, unrenewed pastures in Waikato and Bay of Plenty from spring 2009 to winter 2012, expressed as a percentage of total DM. Data are averaged over both regions.

There was an interaction for sown grasses in 2010–2011, with a higher contribution of sown grass in renewed than unrenewed pastures in Waikato (79% vs 69% of total DM) but a similar contribution in renewed and unrenewed pastures in BoP (averaging 77% of total DM, P = 0.019, Fig. 6). The contribution of unsown grass was higher in Waikato than BoP in 2009–2010 (14% vs 7% of total DM, P = 0.044).

Tiller densities

Tiller densities of perennial ryegrass were less in renewed than unrenewed pastures in spring 2013 (4260 vs 5460 tillers m⁻²). There was an interaction in perennial ryegrass tiller density in autumn 2012 with fewer perennial ryegrass tillers in renewed than unrenewed pastures in BoP (2820 vs 4050 tillers m⁻²) but no effect of renewal in Waikato (averaging 4090 tillers m⁻², P = 0.007). There was no other effect of renewal or region on perennial ryegrass tiller density (4397 ± 554 tillers m⁻² averaged over all years and treatments [± standard deviation]). There were no effects of renewal or region on the tiller densities of unsown grasses (1220 ± 438 tillers m⁻² averaged over all years and treatments [± standard deviation]). Unsown grasses comprised predominantly summer-active C₄ annual grasses such as summer grass (Digitaria sanguinalis [L.] Scop.), smooth witch grass (Panicum dichotomiflorum Michx.), crowsfoot grass (Eleusine indica [L.] Gaertner) or annual poa (Poa annua L.) depending on the time of sampling.

Endophyte presence

There was no renewal or regional effect on endophyte infection frequency in any of the years assessed (84% ± 3% [± standard deviation]). The lowest infection frequency occurred in Waikato in 2012 (69%) and the highest in BoP in 2010 (97%).

Invertebrate species presence and abundance

White fringed weevil (Naupactus leucoloma Boheman) abundance below ground was lower in renewed than unrenewed pastures in 2010, 2011 and 2013 (Table 3). Total invertebrate abundance below ground was lower in renewed than unrenewed pastures in 2010.

Table 3 Density of below-ground (soil) and above-ground (foliar) dwelling invertebrates and earthworms in renewed and unrenewed Waikato and Bay of Plenty (BoP) dairy pastures, assessed annually in February 2009–2013.

		Renewed				Region
Density (number m^−2)^a	Year	Yes	No	SED	P value	Waikato	BoP	SED	P value
Black beetle ^(soil)	2009	26	42	14.1	0.083	21	47	14.1	0.031
White fringed weevil ^(soil)		7	10	6.0	0.236	3	14	6.0	0.080
Earthworms ^(soil)		159	201	44.9	0.135	220	140	77.2	0.216
Clover root weevil ^(soil)		20	12	11.5	0.770	19	13	13.5	0.462
Total invertebrates ^(soil)		95	100	33.1	0.365	85	110	33.1	0.711
Total nematodes ^(soil)		115	116	37.7	0.437	126	104	55.3	0.459
Black beetle ^(soil)	2010	12	38	6.0	0.001	20	29	7.9	0.109
White fringed weevil ^(soil)		3	44	20.0	0.004	7	41	21.3	0.130
Earthworms ^(soil)		156	235	39.0	0.206	218	173	54.3	0.331
Clover root weevil ^(soil)		11	10	5.5	0.920	12	9	5.5	0.674
Total invertebrates ^(soil)		61	155	31.4	0.008	121	95	48.8	0.405
Total nematodes ^(soil)		55	55	5.2	0.879	72	38	29.1	0.326
Clover root weevil ^(foliar)		4	3	0.8	0.219	3	4	1.4	0.535
Argentine stem weevil ^(foliar)		34	20	10.8	0.129	32	23	10.8	0.979
Black beetle ^(soil)	2011	9	12	4.8	0.278	12	9	5.1	0.737
White fringed weevil ^(soil)		1	13	3.5	0.001	5.4	9	3.6	0.259
Earthworms ^(soil)		213	165	93.5	0.719	nd	nd	nd	nd
Clover root weevil ^(soil)		6	4	3.1	0.659	4	6	4.8	0.769
Total invertebrates ^(soil)		44	67	12.1	0.107	65	46	15	0.909
Total nematodes ^(soil)		83	118	12.8	0.023	123	77	22	0.071
Clover root weevil ^(foliar)		1	1	0.6	0.356	1	1	0.6	0.903
Argentine stem weevil ^(foliar)		21	11	6.6	0.023	21	10	8.0	0.110
Black beetle ^(soil)	2012	6	10	3.5	0.786	12	4	3.5	0.048
White fringed weevil ^(soil)		3	7	2.3	0.118	8	1	3.0	0.037
Earthworms ^(soil)		223	254	45.5	0.363	257	219	50.2	0.635
Clover root weevil ^(soil)		45	46	10.5	0.879	63	29	13.7	0.087
Total invertebrates ^(soil)		87	90	20	0.643	126	52	23.5	0.016
Total nematodes ^(soil)		97	83	17.2	0.578	113	67	25.5	0.093
Clover root weevil ^(foliar)		3	2	0.8	0.319	3	1	0.8	0.005
Argentine stem weevil ^(foliar)		5	4	3.3	0.748	4	6	3.3	0.334
Black beetle ^(soil)	2013	5	9	3.3	0.960	9	5	5.6	0.966
White fringed weevil ^(soil)		1	3	0.7	0.048	1	3	1.2	0.183
Earthworms ^(soil)		101	109	49.2	0.987	43	167	49.2	0.041
Clover root weevil ^(soil)		3	7	2.6	0.371	1	9	4.0	0.068
Total invertebrates ^(soil)		35	34	14.4	0.408	39	30	18.4	0.994
Total nematodes ^(soil)		61	70	6.4	0.139	57	74	10.4	0.148
Clover root weevil ^(foliar)		1	1	0.3	0.863	0.4	1	0.3	0.076
Argentine stem weevil ^(foliar)		9	3	4.1	0.323	3	9	4.3	0.226
Black beetle ^(soil)	2014	14	20	6.0	0.382	14	20	15.3	0.576
White fringed weevil ^(soil)		11	18	2.6	0.171	26	3	17.9	0.176
Earthworms ^(soil)		180	155	45.2	0.759	80	255	54.7	0.027
Clover root weevil ^(soil)		11	7	6.4	0.438	11	7	7.3	0.761
Total invertebrates ^(soil)		60	67	17.2	0.518	78	49	26.4	0.532
Total nematodes ^(soil)		63	51	11	0.153	62	52	16.4	0.817
Clover root weevil ^(foliar)		3	2	0.7	0.527	2	3	1.2	0.316
Argentine stem weevil ^(foliar)		9	3	5.1	0.648	3	8	5.1	0.183

^aNematode abundance is per g dry soil; nd, not determined.

SED, standard error of difference.

Black beetle abundance below ground was lower in Waikato than BoP in 2009 but differences reversed in 2012. Similarly, the abundance of white fringed weevil and total invertebrates below ground, and clover root weevil (Sitona obsoletus[= S. lepidus]) abundance above ground was greater in Waikato than BoP in 2012. Earthworm abundance was lower in Waikato than BoP in 2013 and 2014. Clover root weevil abundance was low throughout the sampling period and there was no effect of renewal on their abundance.

Total nematode abundance was lower in renewed than unrenewed pastures in 2011 (Table 3). The abundance of cyst nematodes was greater in renewed than unrenewed pastures in 2012 (0.5 vs 0.2 g⁻¹ dry soil, P = 0.011) and 2013 (2.0 vs 1.5 g⁻¹ dry soil, P = 0.016). In contrast, the abundance of pin nematodes was lower in renewed than unrenewed pastures in 2013 (7.0 vs 16.3 g⁻¹ dry soil, P = 0.043) and 2014 (10.6 vs 14.4 g⁻¹ dry soil, P = 0.048).

There were interactions between renewal and region for lesion and root knot nematode abundance. Lesion nematode abundance was greater in renewed than unrenewed pastures in 2010 in BoP (9.9 vs 3.6 g⁻¹ dry soil) but similar in renewed and unrenewed pastures in Waikato (averaging 7.5 g⁻¹ dry soil, P = 0.011). The abundance of root knot nematodes was lower in renewed than unrenewed pastures in 2011 in Waikato (1.5 vs 5.9 g⁻¹ dry soil) but similar in renewed and unrenewed pastures in BoP (averaging 2.5 g⁻¹ dry soil, P = 0.006).

Soil nutrients and pH

Calcium levels were higher in Waikato than BoP pastures in 2011 (9.0 vs 6.9 me 100 g⁻¹, P = 0.017) and 2012 (10.7 vs 7.0 me 100 g⁻¹, P = 0.049). There were no other significant differences between soil nutrients or pH between renewed and unrenewed pastures or regions (Table 4).

Table 4 Soil nutrient levels in renewed and unrenewed dairy pastures, averaged across region (Waikato and Bay of Plenty) and year (2011, 2012 and 2013).

Soil nutrient	Renewed	Unrenewed	SED	P value
pH	6.0	6.0	0.08	0.951
Ca (me 100 g^−1)	8.5	9.2	0.70	0.333
Olsen P (mg L^−1)	61	58	5.7	0.592
K (me 100 g^−1)	8.6	8.5	1.3	0.943
S (SO4) (mg kg^−1)	20	23	3.4	0.318
Mg (me 100 g^−1)	25	28	3.6	0.333
Na (me 100 g^−1)	7.6	7.2	0.73	0.619

SED, standard error of difference.

Discussion

Production benefits of pasture renewal

Production benefits from pasture renewal were demonstrated by greater herbage production in renewed pastures in winter 2010, summer 2010–2011 and spring 2011. In addition, there was a significant increase of 10%, 11% and 12% in total annual herbage production in renewed pastures for the first, second and third years of monitoring, respectively. These results are similar to those in a Waikato study where there was an 11% increase in pasture production on a farmlet comprising recently renewed pastures (<3 years old) in comparison with production on a farmlet with older pastures (>10 years old), when assessed over 4 years (Glassey et al. 2010).

In the current study, renewal benefits diminished over time such that significant differences in total annual herbage production did not extend beyond the third year. However, there was a consistent trend of greater herbage production in renewed than unrenewed pastures throughout the study. This was reflected in a 13% increase in herbage production over summer (equating to 3.1 t DM ha⁻¹) and an 8% increase overall (equating to 6.5 t DM ha⁻¹), in the renewed pastures when summed over 5 years. The additional feed over summer would assist in reducing the amount of supplementary feed required during this period, which would be particularly valuable during drought years.

There was significant variability in herbage production when comparing renewal effects between farms in the same year and also renewal effects for the same farms between years, as demonstrated by Fig. 5. Some renewed pastures produced less than their paired unrenewed control pastures, particularly in Waikato as the pastures aged. Better performance of pastures in the first few years after renewal may be related to several factors, including a lower content of broadleaf weeds and the lower abundance of insect pests such as black beetle, white fringed weevil and nematodes that feed on ryegrass and clover (Zydenbos et al. 2011).

The lack of significant differences between renewed and unrenewed pastures in sown grass dry matter production is consistent with the tiller density data. Perennial ryegrass tiller populations in renewed pastures rapidly reached equilibrium densities similar to those in the unrenewed pastures. Perennial ryegrass densities reported here are typical of those reported in other New Zealand studies but are lower than those of pastures reported in Europe. For example, compare Laidlaw (2004) who reported densities of 6370 tillers m^-2 under a cutting regime (averaged over six diploid and three tetraploid cultivars) and Tozer et al. (2014) who reported densities of 3030 tillers m⁻² (averaged over 24 grazed dairy pastures of which 18 were sown with diploid and six with tetraploid cultivars). Lower perennial ryegrass tiller densities are associated with poorer performance (Stewart & Hayes 2011) due to impacts on ryegrass population biology (Korte et al. 1985) and are also consistent with the concerns expressed by farmers regarding poor pasture persistence in Waikato and Bay of Plenty (Kelly et al. 2011).

Impacts of renewal on pasture composition and invertebrates

There were some improvements in the botanical composition of renewed pastures with significantly greater clover and lower broadleaf weed herbage production and a greater sown grass contribution for at least one of the first 2 years. For example, nearly twice as much clover was produced in renewed than unrenewed pastures when averaged over the first 2 years (1320 vs 710 kg DM ha⁻¹). Increasing clover content improves pasture quality by increasing metabolisable energy and digestibility, leading to greater livestock productivity. It is estimated that livestock performance is on average 40% higher on clover than on ryegrass and that clover is utilised more efficiently than ryegrass by livestock for growth (Nicol & Edwards 2011).

The greater clover content in the renewed pastures, particularly in BoP, may be explained by the type of perennial ryegrass cultivars sown. The majority of pastures in BoP were sown with the tetraploid cultivar, Bealey. It is thought that tetraploid cultivars are more ‘clover friendly’ than diploid cultivars, with higher clover contents noted in some studies (Stewart & Hayes 2011) but not others (Thom et al. 1998). The greater clover content may also have contributed to the greater abundance of earthworms in BoP than Waikato in 2013 and 2014, as increases in earthworm abundance are associated with increases in clover content (van Eekeren et al. 2008). The increased abundance of clover-feeding cyst nematodes in renewed pastures was also likely to be in response to increases in clover content after renewal (Bell et al. 2009). In turn, the damage these nematodes do to clover roots would subsequently reduce clover growth and nitrogen fixation (Yeates et al. 1977; Yeates 1978). Of the root knot nematode species identified, two (Meloidogyne hapla Chitwood and M. trifoliophila) were most likely feeding on the white clover component of the pasture whereas M. fallax is capable of feeding on both clover and grass components of pasture.

Despite the higher clover content in renewed than unrenewed pastures, the contribution of clover remained relatively low, averaging 10% of DM of all pastures over the duration of the study (Fig. 6). This may be explained by the high rates of nitrogen applied to the Waikato and Bay of Plenty dairy pastures (c. 180 kg N ha⁻¹ yr⁻¹; S. Beynon, BOP farm consultant, unpubl. data) which would have a negative impact on clover growth (Harris et al. 1996). Results are consistent with an upper North Island study in which Harris et al. (1996) reported that clover comprised 10% of DM in dairy pastures to which 200 kg N ha⁻¹ yr⁻¹ had been applied. Changes in management, such as a reduction in nitrogen application, may increase the clover content. However, the economic implications of any management change would need to be considered. Another option may be to oversow clover, although competition with grasses would make successful establishment difficult.

One unexpected observation in this study was the greater unsown grass production in renewed pastures. The most likely reason for this is the disturbance of the soil surface through cultivation and creation of bare ground during the renewal process. This enhances the germination and establishment of seeds present in the soil seedbank, which is dominated by unsown grasses in Waikato and BoP dairy pastures (Bell 1995; Tozer et al. 2010, 2011).

This highlights the importance of controlling grass and also broadleaf weeds during the renewal process to prevent weeds competing with the sown species and reducing the sown species’ performance. The inclusion of a cropping phase, as occurred in half of the renewed pastures monitored in this study, would enable grass and broadleaf weeds to be chemically controlled and prevent their seeds from replenishing the soil seedbank. A further option is to use selective herbicides (James et al. 2013) in a perennial pasture post-establishment to control unsown summer-active C₄ annual grasses such as summer grass that were prevalent in this study. C₄ grasses have significantly lower nutritive value than sown species and thereby reduce pasture quality (Jackson et al. 1996) and livestock production. The increase in herbage production of unsown grasses and especially summer-active grasses in the renewed pastures in this study would therefore offset some of the benefits gained through an increase in herbage production and clover content. Along with inter-plant competition from C₄ plants contributing to a decline in sown species performance, these and other unsown grasses contribute to increasing invertebrate abundance. Unsown grasses such as summer grass and annual poa lack the protection from insect pests that is imparted by endophytes of sown perennial grasses and can act as a favourable food resource for these invertebrates (Zydenbos et al. 2011). Where this occurs, greater populations of some pests build up and can have an impact on both unsown and sown grass performance.

Factors affecting pasture performance

There were major regional effects on pasture performance, with significant differences between regions in herbage production, tiller densities, botanical composition and the abundance of insect pests. Regional differences were often greater than renewal effects (renewed vs unrenewed), inferring that climate is a key driver of differences in pasture performance. The warmer, wetter and sunnier climate of BoP (over the study duration) would favour perennial ryegrass and clover growth, particularly over summer when pastures are most likely to be exposed to heat and moisture stress. Waikato was subjected to a number of summer droughts during the study; rainfall was greater in BoP during these periods. The lower insect pest abundance in BoP is another factor contributing to greater herbage production and a more favourable botanical composition in the BoP pastures assessed.

The high proportion of tetraploid varieties in the renewed BoP pastures may have contributed to the lower tiller densities in renewed compared with unrenewed pastures as occurred in autumn 2012 (by 30%). Lower tiller densities in pastures sown with tetraploid cultivars rather than diploid are consistent with other studies (Orr et al. 2005; Tozer et al. 2014). However, there were no significant differences in tiller densities on other occasions and the trends in renewal effect (Fig. 5) demonstrate that there were definite advantages for pasture renewal on the farms in BoP when compared with Waikato, regardless of the type of ryegrass sown.

Endophyte infection frequency was similar in renewed and unrenewed pastures, and is unlikely to be a reason for the better performance of the renewed pastures. However, other data from this study demonstrated that there was no relationship between endophyte infection frequency and pasture production (Rennie et al. 2011). Soil nutrient levels were also similar in renewed and unrenewed pastures so are unlikely to be a major factor in determining differences between renewed and unrenewed pastures in herbage production or botanical composition. Declining performance over time most likely occurs due to a combination of a number of the above factors as well as abiotic stresses such as drought.

Impacts of region and climate were also evident in a field survey in which the contribution to herbage production and tiller density of perennial ryegrass were far greater under irrigation in Canterbury than in dryland Waikato–Bay of Plenty dairy pastures (Tozer et al. 2014). In the absence of severe moisture stress and insect pest outbreaks, pasture renewal may have fewer benefits. For example in irrigated Canterbury pastures, herbage production in renewed and unrenewed irrigated dairy pastures was similar within 2 years of renewal (Taylor et al. 2012), which is in contrast to this study, where differences between renewed and unrenewed pastures were maintained for longer.

Understanding the factors behind the differences in performance will require comprehensive knowledge of the interaction between the biological drivers of persistence and on-farm management decisions, which is beyond the scope of this study. Reasons for the relatively poor performance of renewed pastures in Waikato compared with Bay of Plenty require particular attention. The Waikato region supplied 23% of New Zealand’s milk solids and contained 24% of New Zealand’s dairy cows in 2012–2013 (DairyNZ 2013), so an ability to lift production via successful pasture renewal is a critical challenge in this region.

Value proposition of pasture renewal

The value to the farmer of the additional feed from pasture renewal will depend on the establishment costs and the value of the feed produced. Project team farmers estimate that the direct costs associated with pasture renewal were approximately $1300 ha⁻¹ (S. McHardy, pers. comm.). If the estimated value of feed produced over the whole year ranges from $0.33–0.39 kg DM ha⁻¹ depending on the economic models used (Chapman et al. 2012) and an additional 6.5 t DM ha⁻¹ were produced in renewed pastures over the 5 year study, renewal would be profitable for most farmers. However, some farmers would make a loss, as demonstrated by renewal effects that show the performance of the renewed pastures in some instances were poorer than the unrenewed pastures on the same farm. This loss may be greater in Waikato than Bay of Plenty, based on the renewal effect data presented in Fig. 5.

The potential for production and economic losses to occur highlights the importance of best practice establishment and management to maximise the production from new pastures. For example, appropriate endophyte selection will help to minimise the negative impact of pests on plant survival (Popay & Hume 2011). The use of break crops during renewal, to break the disease cycle and reduce weed seed abundance and replenishment of the soil seedbank through timely application of herbicide, will also improve establishment and subsequent pasture performance (Bell et al. 2006). Additionally, good grazing practices during drought will assist in sown species persistence and maintenance of renewal benefits, for example, by preventing overgrazing and providing supplementary feed (Clark 2011).

In conclusion, pasture renewal in Waikato and Bay of Plenty is an option to lift pasture performance through increased herbage production and improved botanical composition. However, pasture renewal may be necessary on an ongoing basis, possibly every 5 years, for production benefits to be sustained. Although the benefits of pasture renewal generally declined over time, some renewed pastures consistently performed better than their unrenewed controls over the 5 year study. This demonstrates that pasture renewal has the potential to improve performance and achieve sustained long-term benefits. Key biotic and abiotic factors require investigation to understand the drivers of pasture performance including interactions between drought, insect pest and defoliation stress and perennial ryegrass with different genetics, flowering dates and ploidy. This will require a series of controlled studies that can then be extended to field conditions to better understand interactions with farm management and their impacts on persistence.

Acknowledgements

Thanks to DairyNZ and the Ministry for Primary Industries (through Sustainable Farming Fund: SFF 08-019, 11-089, 11-035) for providing funding for this Bay of Plenty Focus on Dairy initiated project. We are also grateful for the additional support of Ballance Agri-Nutrients, Agricom, PGGW Seeds, Agriseeds, BoP Regional Council and Waikato Regional Council. Many thanks to farmers in Bay of Plenty and Waikato for making their views known and their paddocks available. We thank Linda Trolove and Tina Eden of AgResearch for their assistance with pasture and insect analyses, respectively, and Errol Thom and the referees for their extremely helpful comments on the manuscript.