Effect of nitrogen fertilization, grass species and cultivar on sod production on Valkeasuo peat bog - a case study Perttu Virkajärvi Agricultural Research Centre ofFinland, North Savo Research Station, FIN-71750 Maaninka, Finland, e-mail:perttu. virkajarvi@mtt.fi Harri Huhta Agricultural Research Centre ofFinland, Research Station forEcological Agriculture, FIN-51900 Juva, Finland Hannu Tuuri Agricultural Research Centre ofFinland, Data and Information Services, FIN-31600 Jokioinen, Finland As part of a research project concerning the agricultural utilization of cut-away peat bogs, a sod production experiment was conducted at Valkeasuo, Tohmajärvi, in 1990-1993. The aim of the ex- periment was to study the effect of nitrogen and choice of cultivar on sod production and sod quality on peat bogs. The N fertilization rates were 50, 100 and 150kg ha’ 1. The Poa pratensis cultivars were ‘Conni’, ‘Cynthia’, ‘Haga’ and ‘Julia’, the Festuca rubra cultivars were ‘Center’, ‘Juliska’, ‘Koket’ and ‘Näpsä’ and the Agrostis capillaris cultivar was ‘Rasti’. Two mixtures of P. pratensis!F. rubra and one of A. capillaris/F. rubra imitated commercial sod products. Increasing of N fertilization from 50 kg up to 150 kg ha’ 1 a had positive effect on general the quality of sod as well as on the green cover before and after transplanting. It increased the thatch formation. The positive effect of N on the number of tillers and green cover in the year following transplanting was dependent on the species and the cultivar. Species and cultivar affected all meas- ured variables excluding thatch formation. Generally, the P. pratensis cultivars tested suited better for sod production than cultivars of F. rubra, but there were clear differences between cultivars with- in species as well. Although the soil was infertile, the contents of Ca, K, Mg, P, Cu, Fe, Mn, Mo and Zn in the herbage samples were within normal range. The botanical purity was high, which supports the hypothesis that the absence of seed bank of weeds on peat bogs immediately after harvesting the peat can be utilized. Key words: Agrostis capillaris, Festuca rubra, lawns, peat soils, Poa pratensis ntroduction tares of that area becomes cut-away every year. The after use of the cut-away peat bogs has been an important question since the 1980’s, and sev- eral possibilities instead of abandoning or restor- ing mires have been proposed (Heikkilä 1990). The total area of mires in peat production is about 50.000 hectares in Finland. Nearly 2.000 hec- © Agricultural and Food Science in Finland Manuscript received February 1997 269 Vol. 6 (1997): 269-281. AGRICULTURAL AND FOOD SCIENCE IN FINLAND https://www.c-info.fi/en/info/?token=GaHzIB9UTf-HUHmF.TlPhQJlobikegHxT86G33Q.ttbz-lL20zrj0SRk3zprHOPrZaLrjpN1IAaO-fLK6gpIdFM605xtYZEWhdNUv6Tdd1d5Ej9tYerI9Vje3eXmt67p3RVglkvsoeTem6J23T2OlBzXPdRDVN01T7xQkUtqukAPppU9hMJ4w-DsVmNNqG1pM6iS4pqNFiQgHA1mKbmHp69B5SGlCjvfm9X0W1BAnrDIQp-3g-T2Em1x8oxgnrNyeKpAaWZc063MRiBWRH2rtV0_p5dIpDlLlavSfeEod-21IBKoX8HEwR2z6zaqIHk0s0ETuSaBJx7628UOQth4nQwa7vqW Virkajärvi, P. et al. Effect offertilization, grass species and cultivar on sod production As part of a research project concerning the ag- ricultural utilisation of cut-away peat bogs, a sod production experiment was conducted at Valkeasuo, Tohmajärvi (62%20’N, 30%15’E) in 1990-1993. In sod production the requirements for the soil are high. However, peat bogs do have some advantages which make them potential for sod production. First, immediately after the peat has been harvested, the soil is practically free from weeds and soil-borne diseases. For example, Poa annua, which is generally a difficult weed in lawns, has lowered in some occasions the quali- ty of delivered sod on sport turfs (Erlund 1991, Lahdensuo 1993). Second, some bog areas are practically free from stones, which is a definite requirement for sod production. On the other hand, pieces of undecomposed timber in the soil may prevent sod production on some areas, since they damage the sod during the lifting process. Third, the paddocks on peat bogs are often large and flat. The infertility of the peat bogs must not be a disadvantage: if fertilizing and irrigation are performed properly, the roots will remain at the top layer of the soil.When the sod is then moved to a more fertile substrate, it will root effective- ly because the roots grow rapidly into the new, fertile subsoil. The aim of the experiment was to study the effect of nitrogen and choice of cultivar on sod production and sod quality on peat bogs. Nitro- gen is known to affect many turf quality param- eters, e.g. growth pattern, density, colour, cold- hardiness and botanical composition of the turf community (Beard 1973). High amounts of ni- trogen will accelerate the growth of the tillers at the expense of root growth and thus increase the need ofmowing and litterproduction (Turner and Hummel 1992). On the other hand, nitrogen fer- tilization must be adequate, since nitrogen defi- ciency should neither restrict the density of till- ers nor impair the colour of the turf. In Finnish trials with Poa pratensis and Fes- tuca rubra, it has been shown that cultivar has a major effect on the lawn quality (Huusela-Veis- tola et al. 1991, Niemeläinen and Huusela-Veis- tola 1991). Since peat bogs are rather difficult areas with reference to overwintering conditions, only high- or medium-quality cultivars of P. prat- ensis, Festuca rubra ssp. rubra, Festuca rubra ssp. communtata and Agrostis capillaris were chosen. In addition to pure cultivars, also mix- tures of P. pratensis and Festuca as well as A. capillaris and Festuca were sown to imitate com- mercial sod products. Material and methods Establishing the experiment On the study area, the top soil was coarse sand, the proportion ofhumus being 15 ± 6.8 % (mean ± sd). The area was limed with 7.3 tons per hec- tare limestone containing Mg and fertilized with a compound fertilizer (30 kg N, 21 kg P, 42 kg K ha 1 )- Soil analyses after lifting the sod showed the following concentrations of soluble nutrients (mg I 1): Ca 933, K 39, Mg 56 and P 1. 1. The pH (H,O) was 5.4. The plot size was 300 * 45 cm with 25 cm zones at both ends of the plots for sampling procedures. The plots were sown by hand on 24 August 1990. The experiment had threerep- licates. The treatments are presented in Table 1. The plots were fertilized with 10 kg N, 50 kg P and 90 kg K ha’ 1 (as compound fertilizer in- cluding trace elements) in May 199land 1992. One half of the nitrogen doses was applied si- multaneously while the other half was applied at the end of June. The lawn was mown to the height of 3-4 cm twice a week during early sum- mer and once a week during late summer. In the autumn 1990 and 1991, Quintozene (13-16 kg ha 1 Avicol) was sprayed against fungi (e.g. Fusarium nivale, Typhula sp. Sclerotinia borealis) to prevent overwintering damage. Observations and measurements The establishment was estimated on the basis of the green cover (0-100%). In the first growing year the green cover was estimated five times 270 AGRICULTURAL AND FOOD SCIENCE IN FINLAND and in the lifting year and the following year once a week. The shoot density was counted from core samples (0 5.0 cm), two per plot in 1991 and 1992. The root mass was determined from sub- plot treatments 2,5, 8 and 10 by taking core sam- ples (0 7.2 cm) to the depth of 12 cm. The sam- ples were cut horizontally into five layers: 0-2.0, 2.0- 4.5, 4.5-7.0, 7.0-9.5 and 9.5-12.0 cm. The samples were stored at a temperature of -1 to O°C and taken to room temperature for 24 hours before washing on a sieve. Subsequently, they were oven-dried(24 hours at 100°C). Before lifting, the quality of the sod was es- timated as follows: Characteristic Scale Colour 0-5 (0 = defect in colour) Density 0-5 (0 = sparse) Uniformity 0-5 (0 = incoherent) Weeds 0-1 (0 = weeds disturb) The sum variable of sod quality (maximum 16 points) was taken as a continuous variable. Samples of roots and tillers were taken again on 11 August 1992 before lifting. A bulk sample of herbage of P. pratensis and F. rubra was cut to the height of 2 cm to estimate possible nutri- ent deficiency. N, P, K, Ca, Mg, Cu, Fe, Mn, Mo and Zn were analysed from the samples in the Central Laboratory, Agricultural Research Cen- tre, Jokioinen. 250*45 cm2 pieces of the sod were lifted with the machinery ofKarjalan Nurmi Ltd on 14 Au- gust 1992. A 150*45cm 2 piece of each plot was assembled in the same order as before on miner- al soil (pH 6.1, Ca 1240, K 40, P 9.9 mg/1). The area was fertilized (30 kg N, 21 kg P, 42 kg K ha 1) and well irrigated before transplanting. The trial was carefully rolled after assembling. The handling tolerance of sod was measured as tearing force from three samples (21*21 cm) per plot as described by Shildrick (1982). The rooting of the sod was evaluated five weeks af- ter transplanting by core samples (0 5.0 cm), two per plot, and given values 0-5 (0 = not rooted at all, sod loose from subsoil, no new roots visi- ble; 5 = sod well rooted, removing from subsoil difficult, new roots abundant in subsoil). Table I. Treatments. Main plot: Nitrogen fertilization (kg ha' 1 year 1 ). 1: 50 2: 100 3: 150 Subplot: species, cultivar or mixture No Species Cultivar Seed rate g nr2 Cultivar 1 P. pratensis Conni 10 2 “ Cynthia 3 “ Haga 4 “ Julia 5 A. capillaris Rasti 8 6F r. communtata Center 20 7 F. r. rubra Juliska BFr communtata Koket 9 F.r. rubra Näpsä 10Mixture 1° Center/Juliska// Conni/Julia/Haga 16 11 Mixture 2I)3) Center/Wilma// Conni/Julia/Cynthia 16 12Mixture 3 21 Rasti/ZCenter 10 11 20 % each cultivar 2) 80 % A. capillaris, 20 % F. rubra 31 Wilma =Ere. The botanical composition of the plots was observed on 28 June 1993 with a point quadrat, 100 points per plot. The observations were di- vided into six classes: F. rubra, P. pratensis, A. capillaris, other monocotyledons, dicotyledons and gaps. Statistical Methods The data was analysed mainly by the analysis of variance according to the split-plot design. Rep- lication was considered a random effect where- as N fertilization as well as species/cultivar were considered fixed effects. N fertilization was analyzed as main plot factor and species/cultivar as the subplot factor. Analyses were performed by means of the SAS statistical package. MIXED (SAS 1992) and GLM (SAS 1991) procedures were used. 271 Vol. 6 (1997): 269-281. AGRICULTURAL AND FOOD SCIENCE IN FINLAND Virkajärvi , P. et al. Effect offertilization, grass species and cultivar on sodproduction In analysing root mass the years and depths were analysed as repeated-measures factors in the same split-plot experiment mentionedabove. The multivariate approach was used in the anal- ysis and it was performed by means of the GLM procedure of the SAS. After this the most im- portant layer (0-2 cm), was examined separate- ly. In these analyses year was the only repeated- measures factor. Green cover data for the lifting year (1992) and the following year (1993) were analysed separately. In both years the growing season was divided into three periods. The green cover val- ues for each period were already means of six weekly observations done in each period. Thus the repeated-measure factor, period, had three levels. Green cover was analysed using the mul- tivariate analysis of variance with period as a repeated-measures factor in the split-plot exper- iment. In the sequel, periods were examined separately because the effects of nitrogen fertiliza- tion and species were different in differentperiods. Rooting ability, which is an ordinal-scaled variable, was analysed using a split-plot, non- parametric analogue. The analysis was based on recommendations by Koch et al. (Koch 1970, Koch et al. 1980) regarding the non-parametric approach for repeated measurements. All effects were tested separately using the Friedman test and by means of StatXact 3 for WINDOWS soft- ware (CYTEL Software Corporation 1995). The testing of interaction effect between N fertiliza- tion and species/cultivar was divided into three parts because of the three levels of N fertiliza- tion (main plot factor). Therefore, when the sig- nificance of interaction was examined, the number of tests was taken into account by using Bonferoni’s approach (Rosenthal and Rosnow 1985, p. 45). Prior to the analysis, however, the data had to be pooled in several ways to allow tests of various effects. All these modifications were doneby means of the SAS. The botanical composition was examined in the final analysis as a three-compositional fac- tor (sown species, gaps and others).The propor- tions that describe the botanical compositions sum up to 1 over all botanical types. So, they are obviously not independent. It can therefore be misleading to analyse the data for each botani- cal type separately. Our strategy for the analy- ses was to convert the compositions to additive log-ratio compositions, to reformulate the under- lying problem in terms of these log-ratio com- position and to make use of the multivariate anal- ysis of variance (Aitchison 1986). The GLM pro- cedure of the SAS was used for analyses. Be- fore calculating the log-ratio compositions, 0% in the botanical composition was assigned 0.1 %. In general, it was not found necassary to ex- amine pairwise differences. However, if multi- ple-comparison procedures were needed, Tuk- ey’s HSD method or t-type contrast examination was used. Before performing the analysis of variance, accordances of data with the distributional as- sumptions of the models were checked by diag- nostic methods (e.g. Box-Cox diagnostic plots) on what basis some variables were transformed, the germinating percentage was square root- transformed, theFe content 1/x-transformed, the Cu and Mn contents log(x)-transformed and the root mass log(l+x)-transformed. Results Germination and establishment were satisfacto- ry. This was confirmed by the high green cover values in 1991. The weather data are presented in Table 2. In the lifting year of 1992, the effect of ni- trogen on the green cover was different in dif- ferent periods (p=0.030 for interaction of N, SP and Period). Therefore, analyses were carried out by periods. The N fertilization did not affect the green cover in the spring, but it increased the green cover in midsummer (Table 3). However, the magnitude of the effect remained low. After lifting and transplanting the sod the green cover values dropped clearly. This decrease was most sharp with low N and only slight with high N. No interaction with N fertilization and cultivar/ 272 AGRICULTURAL AND FOOD SCIENCE IN FINLAND mixture was detected during any period, which indicates that the effect of N fertilization was almost the same for each species/cultivar. Species/cultivar played a major role in the green cover values in each period. The ranking order between species/cultivar did change in the course of the season. After transplanting, espe- cially ‘Center’, ‘Koket’ and ‘Näpsä’ had low values, while ‘Conni’ maintained a high green cover. The general quality of sod improved by in- creasing N fertilization without interaction with species/cultivar (Table 3). Again, there were clear differences between cultivars. Generally, culti- vars of P. pratensis and A. capillaris ‘Rasti’ had higher values than F. rubra cultivars. The par- tial rankings (density, colour, uniformity, pro- portion of weeds) had the same pattern as the sum variable, except disturbance ofweeds, which was not affected by N fertilization. In the botanical analysis carried out in the summer following transplanting, the mean pro- portion of gaps was 8% (SD 9.4%). The mean proportion of weeds was only 5% (SD 11%) for all species/cultivars and only 2% (SD 3.3%) when ‘Näpsä’ was excluded. ‘Näpsä’ was a clear exception and it was left out from the analysis of variance. Proportion of sown species, the most important factor in the botanical composition, is described in Table 3. The effect of N fertiliza- tion on the proportion of sown species was not statistically significant. However, there were clear differences between cultivars (also with- out ‘Näpsä’). The proportion of sown species was highest in ‘Rasti’ and lowest in ‘Cynthia’ and ‘Näpsä’, as mentioned above. The low propor- tion of sown species in ‘Näpsä’ was mainly due to its low rooting ability. Its proportion was 35% with 50 kg N and 61% with 150 N. Nitrogen did not affect the rooting, but spe- cies/cultivar did. ‘Näpsä’ rooted most poorly while ‘Cynthia’, ‘Haga’, ‘Julia’, ‘Koket’, Mix- ture 2 and Mixture 3 were a quite homogeneous, well-rooting group, the rest being intermediate (Table 3). The effect of species/cultivar on the number of tillers was clear (pcO.001) in 1992. The ef- Table 2. Monthly mean temperatures and precipitation sums during growing seasons of 1990-1993 in Tohmajärvi. 1990 1991 1992 1993 1961-1990 Mean temperature, °C May 7.8 7.6 9.5 11.0 8.6 June 12.2 13.3 14.5 10.7 14.0 July 15.1 16.1 14.8 15.2 15.9 August 13.7 14.5 13.3 12.8 13.5 September 6.8 8.0 11.3 4.2 8.3 Precipitation, mm May 34 41 18 23 36 June 28 117 40 89 57 July 82 124 54 143 70 August 56 130 105 131 80 September 17 80 46 44 65 feet of N was dependent on the cultivar (N * SP interaction, p=0.032). On average, the number of tillers increased slightly with increasing N fertilization: 41100, 48900 and 54600 tillers nr2 with 50, 100 and 150 kg N ha ', respectively (p=o.ooB, SEM 1236). The effect of subplot was not only due to different species but there were clear differences between cultivars as well. The highest density was observed in ‘Rasti’, 77100 tillers nr2 with 150 kg ha 1 N, and lowest in ‘Haga’, 21900 tillers nr2 with 50 kg ha 1 N. Among F. rubra, the highest observed density was found in ‘Center’ plots, 68500 tillers m 2 with 150 kg ha' 1 N. Nitrogen fertilization increased the thatch formation (p=0.047). The thicknesses of the thatch layers were 12.4, 14.7 and 18.7 mm (SEM 0.65 mm) with 50, 100 and 150 kg N ha 1 , re- spectively. The effect of N fertilization on the tearing force was dependent on the species/cultivar (Ta- ble 4). Nitrogen could either increase or decrease the tearing force. However, the magnitude of the effect was not high, ca ± 10 kg at the most. Spe- cies/cultivar had a much greater effect: the sam- ples of ‘Conni’ tolerated over twice the force that ‘Rasti’ tolerated. In addition to ‘Conni’, ‘Cyn- 273 Vol. 6 (1997): 269-281. AGRICULTURAL AND FOOD SCIENCE IN FINLAND Virkajärvi, P. el al. Effect offertilization, grass species and cultivar on sodproduction Table 3. Effect of nitrogen fertilization and species/cultivar on the green cover, quality, rooting and propor- tion of sown species of sod. Main effect means, standard errors of means (SEM) and significances of F values in the analysis of variance (p values) are presented. Green cover % Quality 1 Rooting Proportion of (Ranked 2 ) sown species’ Spring Summer Transplanted % Nitrogenfertilization (N) N 50 N 100 N 150 SEM Species/cultivar(SP) Conni Cynthia Haga Julia Rasti Center Juliska Koket Näpsä Mixture 1 Mixture 2 Mixture 3 SEM Source ofvariation N SP N*SP '5 53 5.82.0(1.3) 79 12 67 9.22.4 (2.0) 87 15 80 11.32.7 (2.6) 92 0.61.5 0.34 75 82 85 11.7 2.2(5.6) 95 10.2 2.9(8.2) 77 9.6 2.8 (8.3) 89 9.5 2.7(8.1) 89 10.9 2.3(6.3) 96 5.5 2.1(5.3) 85 7.4 2.1(4.6) 90 7.3 2.7(8.2) 83 8.5 0.7(1.2) 49 7.9 2.1(5.0) 94 8.3 2.9 (8.8) 93 8.4 2.8 (8.4) 93 80 64 67 66 67 59 65 61 62 89 83 85 84 91 73 75 73 80 687941 697839 687938 0.582.01.11.1 P values 0.39