Effects of Organic Fertilizers on Soil Physicochemistry and on the Yield and Botanical Composition of Forage over 3 Years

ABSTRACT

Organic wastes have been reported to reduce saturation of the exchange complex by Al in Al-rich acid soils. For 3 years, the main soil fertility properties were studied in plots sown with mixed pasture species. These plots were fertilized with cattle slurry, dairy sludge (DS), or granulated broiler litter (BL) in comparison with mineral fertilizer. Al saturation levels were low after the initial inorganic liming treatment (19.00–33.71%) but tended to rise under all treatments (21.09–61.37%) except BL (8.45–30.98%), which was also associated with the highest average soil pH and the highest average levels of exchangeable Ca²⁺, Mg²⁺, and K⁺. Treatment DS performed similarly to mineral fertilizer in most respects, but it led to greater available P levels. Under the dry conditions of the second and third years of the study, BL and DS treatments were associated with significantly greater forage yields than the other treatments. Under DS treatment, available P levels were too low to allow the maintenance of mixed pasture, clover being eliminated by the less P-dependent species.

IMPLICATIONS

This paper presents important information for the enterprises that produce these types of organic waste. They could use these residues as new resources to achieve a new profit in two ways: sell the product (BL) or save money (not to pay to an environmental management company to apply a landfill program under actual European Union [EU] regulations) and to help some of them to achieve the Environmental Management Certificate (dairy industry) and be under EU environmental law.

INTRODUCTION

Acidification is the main cause of soil degradation in tropical and temperate regions.1 Some 30% of ice-free land is occupied by acid soils,2 and this percentage is growing because of human activities (e.g. use of fertilizers) or their consequences (e.g., acid rain). Paradoxically (in view of the influence of fertilizers), the main agricultural limitation of acid soils is their low pH, partly because it reduces biological activity and nutrient availability but mainly because it increases the solubility, and hence the toxicity and other adverse effects, of Mn and Al.2–4 In acid soils, Al also dominates the exchange complex, favoring the leaching of Ca, Mg, and K and the fixation of phosphate and sulfate anions. Because of the predominant role of Al in the poor fertility of acid soils, which is better predicted by Al-based measures than by soil pH,⁵ current liming practices tend to aim at preventing or counteracting the adverse effects of this cation rather than at instituting neutral pH per se, especially where liming is expensive or soils are highly buffered, because significant improvements in crop yield can be achieved without the attainment of neutral pH.6–8

Although the major commercially available acidity amendments are inorganic oxides, hydroxides, carbonates, and silicates of Ca or Mg derived from limestone, there is increasing interest in acidity amendment as an outlet for agricultural or agroindustrial wastes such as manure and water treatment sludge.9 Several studies have evidenced the efficacy of these materials in neutralizing the acidity of acid soils and subsoils10,11 and thereby increasing forage yield.12–14 In Europe, the agricultural use of organic waste is regulated by Council Directive 91/676/CEE15 concerning the protection of waters against pollution caused by nitrates from agricultural sources; that of sewage sludge by Council Directive 86/278/CEE16; and that of other organic wastes by Regulation 2003/2003 relating to fertilizers17 and Regulation 1774/2002 on animal byproducts not intended for human consumption,18 which contain heavy metals and pathogens.

According to a relatively recent review,19 the concentration of phytotoxic Al in the soil solution is reduced by organic waste through complexation by soluble humic substances, aliphatic organic acids, and solid-phase organic matter and by precipitation because of raised pH. However, most of the papers reviewed involucre in laboratory studies or short-term field experiments. It is also of considerable interest to understand the long-term effects of organic waste treatments on acid soils, crop yield, and the environment.

This paper reports the results of a field experiment in which, after initial inorganic liming, a previously wooded acid soil was treated over 3 yr with mineral fertilizer or with three kinds of organic waste—cattle slurry (CS), dairy sludge (DS), or broiler litter (BL)—as fertilizer/amendment. During this period, the evolution of the physicochemical parameters of the soil, forage yield, and botanical composition was studied under the various treatments.

MATERIALS AND METHODS

Study Site and Weather

The soil studied was a humic Umbrisol20 in the northwest Spanish locality of Goiriz, Lugo (43°19′ north 7°37′ west), which during the period 1971–2000 had an annual mean temperature of 11.5 °C and a mean annual rainfall of 1084 mm that fell mainly in the autumn (35%), followed by winter (29%) and spring (22%).21 Rainfall and temperature during the study period are shown in Figure 1a. Before the conversion of the study site to meadow, its predominant vegetation consisted of trees (Pinus pinaster, Castanea sativa), bushes (Ulex sp.), and ferns (Pteridium aquilinum), and its main pedological characteristics were high organic matter content (9.33%), low pH (pH_H2O 5.5), low P content (P_Olsen, 5.81 mg kg−¹), strong P retention (pH_NaF >10), satisfactory K content (156 mg kg-¹), and high Al saturation (38%). In September 2001, study area vegetation was removed with a view to its conversion to meadow, and because meadow requires Al saturations less than 20%,22 it was treated with 3 t/ha of powdered limestone with a calcium oxide (CaO) content of 60%. It was then divided in four blocks, each comprising five 3- by 1.3-m plots separated by paths 1.65 m wide, and each plot was sown with three species commonly planted for hay meadows in this region of Spain—Lolium perenne L. “Tove,” L. hybridum Hausskn. “Texy,” and Trifolium repens L. “Huia” at densities of 40, 20, and 6 kg ha⁻¹, respectively.

Figure 1. Monthly precipitation and monthly mean temperature in the study area during the study.

Plot Treatments

Five fertilization treatment regimens were defined: treatment M was based on mineral fertilizers; treatments of CS, DS, or BL; and a control treatment of minimal doses of mineral fertilizer (sufficient to ensure that plots were not totally overrun by weeds; see below). CS was obtained from a nearby farm and was applied as a spray in simulation of large-scale practice in the region; DS was obtained from a local dairy as the product of the treatment of waste fluids with active sludge and was applied in the same way as CS; and dried, granulated BL23 was purchased from a local chicken farm. Relevant characteristics of these agroindustrial waste products are listed in Table 1. Each treatment regimen was randomly assigned to one plot in each of the four plot blocks defined in the study area.

Table 1. Mean characteristics of the organic wastes used (n = 10) and their dosage

Organic Waste	Annual Dose	Dry Matter (g L^−1)	pH	Electrical Conductivity (dS m^−1)	C (g kg^−1)	N (g kg^−1)	P (g kg^−1)	K (g kg^−1)	Na (g kg^−1)	Ca	Mg (g kg^−1)	C/N	C/P
CS	50 m^3 ha^−1	18.2	7.1	4.0	400	51	20	96	24	8	7	7.8	20.0
DS	120 m^3 ha^−1	20.0	7.1	3.4	356	62	21	11	32	22	4	5.7	17.0
BL	4500 kg ha^−1	89.1	7.9	11.1	368	40	16	28	16	19	7	9.2	23.0

N was the nutrient used to determine fertilizer application rate.

Treatment M consisted of preliminary application, immediately before sowing, of 300 kg ha⁻¹ of NPK 5-15-13 followed by yearly maintenance doses of N (30 kg ha⁻¹ as ammonium nitrate, 20.5% N) and phosphorus pentoxide (P₂O₅; 45 kg ha⁻¹ as super phosphate, 18% P₂O₅). These yearly maintenance doses were applied in two equal lots in March and in May (in the latter case after mowing and soil sampling, concerning which see below).

Treatment CS consisted of preliminary application of a quantity of CS containing the same amount of N as the preliminary application of treatment M, followed each year by an average of 50 m³ ha⁻¹ of CS in two lots (March and May); minor deviations from 50 m³ ha⁻¹ depended on the composition of the currently available slurry.

The preliminary treatment in treatment DS consisted of a quantity of DS containing the same amount of N as the preliminary application of treatment M, together with sufficient potassium sulfate to make applied K up to the same level as in preliminary treatment M (DS has low K content; see Table 1). This preliminary treatment was followed each year by an average of 120 m³ ha⁻¹ of DS in two lots (March and May); minor deviations from 120 m³ ha⁻¹ depended on the composition of the currently available sludge and the current nutrient status of the soil.24

Treatment BL included no preliminary treatment, consisting only of single annual applications of 4500 kg ha⁻¹ of BL, in March, with the expectation being that in the course of the year approximately 60% of its N content would become available to plants.25,26

The control treatment was similar to treatment M except that it included no preliminary application, and all maintenance applications of N and P were one-third of those used in treatment M.

All fertilizers were applied in accordance with established guidelines27 on the basis of the annual nutrient requirements of mixed meadow: 200, 150, and 150 kg ha⁻¹ of N, P, and K, respectively.28

Sampling

Depending on weather conditions, plots were harvested 2 or 3 times a year, in May (2002, 2003, and 2004), July (2003 and 2004), and/or November (2002 and 2004). Plants were not harvested in November 2003 because of the dry conditions in summer (Figure 1). The mower had a swath width of 1.30 m and cut at a height of 5 cm. On each occasion, the entire crop of each plot was weighed in the field, after which 500- to 1000-g samples were weighed fresh, oven-dried at 70 °C for 48 hr, and re-weighed to enable calculation of dry matter production in kilograms per hectare.

Chemical Analysis

The forage composition was characterized on a dry-weight basis by separating green grass, green clover, other green plant material, and senescent material; oven-drying each category separately; and weighing.

Soil samples were taken in March (before the March application of fertilizer) and after each mowing. The sub-sample of air-dried soil was grounded through a 2-mm screen, and the following determinations were done: pH in water (using 1:1 v/v suspensions); pH in potassium chloride (KCl; using 1:2.5 v/v suspensions); total C and N contents (in a LECO CNS2000 analyzer); sodium bicarbonate (HNaCO₃)-extractable P29 using a Jenway 6300 spectrophotometer; and ammonium chloride (NH₄Cl)-extractable exchangeable cations30 using a Varian 220 spectrometer in atomic absorption mode for Ca²⁺, Mg²⁺, and Al³⁺ and atomic emission mode for K⁺ and Na⁺. Effective cation exchange capacity (CEC_e) was calculated as the sum of Ca²⁺, Mg²⁺, Al³⁺, K⁺, and Na⁺, and Al saturation as 100 × Al³⁺/CEC_e.

Statistical Analysis

For each soil property a two-way analysis of variance was performed with sampling date and treatment as factors and with the date × treatment cross-term included in the model after verification of distributional normality by the Kolmogorov–Smirnov test and of between-group homoscedasticity by Levene's test.31 Because for most properties the cross-term turned out to be statistically signifi-cant (see Results and Discussion), one-way analysis of variance was performed at each sampling date to detect differences between the effects of different treatments. Posthoc comparisons among groups were then performed using least significant difference tests if the corresponding variances were homogeneous and Games–Howell tests for nonhomogeneous variance, in both cases at the 5% significance level. Pearson correlations were calculated and their significance was evaluated. All statistical calculations were performed using SPSS version 15.0.32

RESULTS AND DISCUSSION

Soil Properties

Soil property values varied widely from year to year, in keeping with which the date factor was significant (P < 0.001) in all of the two-way analyses of variance (Table 2). The treatment factor was also significant in all cases (P < 0.01 for K⁺, P < 0.001 in all others), and the effects of date and treatment interfered with each other to a significant extent in all cases except those of pH_KCl, Al³⁺ saturation, CEC_e, and Ca²⁺ (Table 2). The only variable for which subsequent one-way analysis of variance did not detect significant between-treatment differences on at least 50% of sampling dates was CEC_e, for which no such differences were ever detected despite the treatment factor having emerged as significant in the corresponding two-way analysis of variance.

Table 2. Mean values of soil parameters for each sampling date and treatment (n = 4) with analysis of variance results

	2002			2003			2004
Parameter	March	May	November	March	May	July	March	May	July	November	Date	Treatment	Date' Treatment
pH H2O
C	5.61	5.04 a	5.83	5.18	4.96 a,b	5.70 a	5.11 a,b	5.38	5.20 a	5.41
M	5.77	4.90 a	6.07	5.45	5.31 c	5.85 a,b	5.18 b	5.31	5.26 a	5.41	***	***	***
CS	5.70	4.90 a	5.97	5.14	4.89 a	5.59 a	5.04 a	5.30	5.27 a	5.45
DS	5.74	4.92 a	6.10	5.27	5.14 b,c	5.67 a	5.17 b	5.32	5.32 a	5.43
BL	5.60	5.60 b	6.08	5.52	5.67 d	6.06 b	5.17 b	5.35	5.66 b	5.33
pH KCl
C	4.23 a	4.38 a,b	4.76 a	4.43	4.11 a	4.89 a	4.17 a	4.25 a	4.27 a	4.18 a
M	4.40 a	4.32 a,b	4.84 a	4.65	4.55 a,b	5.12 a,b	4.26 a,b	4.37 a,b	4.40 a	4.29 a	***	***	NS
CS	4.39 a	4.25 a	4.83 a	4.62	4.12 a	4.96 a	4.23 a,b	4.40 b	4.37 a	4.24 a
DS	4.39 a	4.39 a,b	4.95 a	4.55	4.30 a,b	4.95 a,b	4.28 a,b	4.35 a,b	4.38 a	4.19 a
BL	4.67 b	4.67 b	5.42 b	4.68	4.62 b	5.31 b	4.39 b	4.48 b	4.72 b	4.49 b
Al saturation (%)
C	33.71 b	23.78 a	51.43 a	40.51 b	61.37 b	37.59 b	52.87 b	49.00 b	32.79 b	57.53 b
M	22.49 a,b	30.55 b	37.39 a	27.32 a,b	27.41 a	21.09 a,b	41.47 a,b	40.16 a,b	25.07 b	41.94 b	***	***	NS
CS	17.77 a	39.98 c	37.47 a	26.21 a,b	52.78 b	37.14 b	45.65 a,b	34.39 a,b	26.50 b	46.37 b
DS	33.32 b	24.15 a,b	40.12 a	32.71 a,b	29.51 a	24.32 a,b	38.72 a,b	36.34 a,b	21.21 a,b	40.58 a,b
BL	19.00 a,b	19.92 a	30.98 a	16.76 a	19.05 a	8.45 a	30.56 a	27.20 a	5.36 a	21.56 a
CECe (cmolc kg^−1)
C	3.97	3.82	3.58	5.43	4.02	4.66	4.42	4.77	5.04	3.67
M	4.82	4.31	3.69	5.88	5.57	5.75	4.63	5.03	5.50	3.94	***	***	NS
CS	4.77	4.32	3.63	5.29	3.90	4.40	4.39	4.93	4.68	3.56
DS	6.00	4.73	4.08	5.87	5.11	5.36	4.86	5.36	5.15	4.17
BL	5.08	5.06	4.49	7.92	4.99	5.61	4.43	5.03	6.39	4.30
Ca^2+ (cmolc kg^−1)
C	1.95	1.88 a	0.87 a	2.28 a	1.10 a	2.01 a	1.40	1.75 a	2.41 a	1.02 a
M	3.03	2.02 a	1.33 a	3.36 a,b	3.66 a,b	3.92 a,b	2.03	2.31 a,b	3.26 a,b	1.65 a,b	***	***	NS
CS	3.11	1.68 a	1.08 a	2.89 a	1.34 a	2.05 a	1.79	2.49 a,b	2.62 a	1.44 a,b
DS	2.84	2.35 a,b	1.35 a	3.10 a,b	2.93 a,b	3.08 a,b	2.25	2.51 a,b	3.20 a,b	1.91 b,c
BL	3.04	3.04 b	2.30 b	5.11 b	3.24 b	4.01 b	2.16	2.64 b	4.04 b	2.54 c
Mg^2+ (cmolc kg^−1)
C	0.18 a	0.31 a	0.24	0.27 a	0.16 a	0.31 a	0.25 a	0.25 a	0.36 a	0.16 a
M	0.25 a,b	0.32 a	0.35	0.35 a	0.37 a,b	0.39 a,b	0.28 a	0.32 a,b	0.43 a	0.24 a	***	***	***
CS	0.13 a	0.24 a	0.34	0.31 a	0.18 a	0.25 a	0.23 a	0.28 a	0.30 a	0.17 a
DS	0.44 a,b	0.49 b	0.44	0.37 a,b	0.42 a,b	0.43 b	0.34 a,b	0.40 a,b	0.41 a	0.26 a,b
BL	0.47 b	0.47 b	0.43	0.69 b	0.29 b	0.69 a,b	0.47 b	0.55 b	1.02 b	0.41 b
Na^+ (cmolc kg^−1)
C	0.20	0.19 a	0.30 a	0.45 a	0.13 a	0.19	0.28	0.25 a	0.29 a	0.15
M	0.20	0.20 a,b	0.29 a	0.43 a	0.14 a	0.16	0.26	0.26 a	0.28 a	0.20	***	***	***
CS	0.23	0.19 a	0.56 b	0.48 a	0.15 a	0.22	0.26	0.28 a,b	0.32 a	0.14
DS	0.24	0.22 a,b	0.32 a	0.46 a	0.19 b	0.31	0.28	0.33 b	0.34 a,b	0.18
BL	0.24	0.24 b	0.18 a	0.60 b	0.21 b	0.21	0.27	0.23 a	0.39 b	0.18
K^+ (cmolc kg^−1)
C	0.35 a	0.54 b	0.32	0.24	0.17	0.39 b	0.16 b,c	0.19 a,b	0.34 a,b	0.21 a,b
M	0.35 a	0.46 b	0.33	0.26	0.18	0.28 a,b	0.14 a,b	0.18 a	0.22 a	0.19 a	***	**	***
CS	0.47 a,b	0.48 b	0.29	0.22	0.17	0.25 a	0.14 a,b	0.19 a,b	0.28 a	0.18 a
DS	0.60 b	0.52 b	0.30	0.21	0.17	0.27 a,b	0.13 a	0.20 a,b	0.19 a	0.19 a
BL	0.39 a,b	0.33 a	0.23	0.33	0.30	0.24 a	0.18 c	0.24 b	0.59 b	0.26 b
P (mg kg^−1)
C	6.30 a,b	6.78 a	7.63 b	4.62 a,b	3.85 a	5.40 a	5.60 a	5.28 a	11.80 a	3.14 a,b
M	8.48 a,b	11.29 b	7.58 b	4.40 a	4.57 a,b	6.67 a	6.03 a	6.24 a	14.38 a,b	2.91 a	***	***	**
CS	8.63 a,b	6.32 a	9.04 b	4.26 a	4.30 a,b	6.25 a	5.97 a	5.07 a	8.08 a	3.00 a
DS	9.09 b	10.37 b	7.60 b	6.25 a,b	5.85 a,b	16.98 b	6.61 a	8.95 b	12.54 a,b	5.32 c
BL	5.73 a	12.45 b	5.62 a	6.61 b	6.35 b	7.81 a	5.83 a	7.48 a,b	16.62 b	4.61 b,c
Notes: Analysis of variance factors. For each parameter data labeled with different letters differ significantly (P < 0.05) for date or treatment. Date' Treatment indicates statistical interaction (significant or no) between these 2 factors.

The value of pH_H2O peaked in November 2002 (which is attributable to the mineral liming in October 2001), in July 2003, and in July or November 2004; and for all treatments except BL pH_H2O had minima in May 2002, May 2003, and March 2004 (BL plots maintained their initial pH in May 2002 and in 2003 showed a minimum in March rather than May) (Table 2). However, under all treatments except BL, there was a clear decrease in within-year variation, from approximately 1 pH unit in 2002 to 0.4 units or less in 2004, and by the end of the study pH values were rather lower than those recorded at the beginning. Under all treatments, pH_H2O was between 0.43 and 0.66 units lower in March 2004 than in March 2002 and pH_KCl up to 0.28 units lower, whereas between November 2002 and November 2004 pH_H2O fell by 0.42–0.75 units and pH_KCl by 0.55–0.93 units, although pH_H2O was approximately 0.4 units higher in May 2004 than in May 2002 under all treatments except BL. Although BL was thus the only treatment associated with a rise in pH at all times of year between the beginning and end of the study, BL plots nevertheless almost invariably had the highest pH (doubtless because of the higher pH of the chicken manure itself; see Table 1), and control or CS plots the lowest. In particular, the pH_KCl of BL-treated plots was significantly higher than that of control plots at all times except May 2002 and March 2003. Different authors reported that the treatments with poultry litter similarly resulted in a higher pH than treatment with commercial fertilizers.33–35

The variables pH_H2O and pH_KCl were the only two soil variables with which Al saturation correlated significantly (P < 0.05), with |r| > 0.8, on at least three sampling dates. In keeping with the results of laboratory experiments,36 on all occasions on which these Al-pH correlations were statistically significant, the correlation was negative (Table 3), as is illustrated by comparison of the data for pH and Al saturation (Table 2). Global negative correlation is also illustrated by the coincidence of pH_KCl maxima with Al saturation minima, and vice versa, in 2003 and 2004, and by the fact that under all treatments except BL Al saturation was on average slightly higher in 2004 than in 2002. However, only BL-treated plots maintained, from March 2003 on, Al saturation levels that were always significantly lower than those of control plots (Table 2). Accordingly, and in keeping with the findings of Berek et al. 37 and Hue12, the absolute levels of exchangeable Al³⁺ were also lower under BL than under other treatments during this period (results not shown).

Table 3. Simple linear correlation coefficients between Al saturation and the variables pH_H2O, pH_KCl, and crop dry matter (DM) content (n = 20 for each date)

Variable	March 2002	May 2002	November 2002	March 2003	May 2003	July 2003	March 2004	May 2004	July 2004	November 2004
pHH2O
r	−0.601	−0.609	−0.310	−0.786	−0.809	−0.852	−0.635	−0.157	−0.885	−0.090
P ^a	0.005	0.004	0.184	< 0.001	< 0.001	< 0.001	0.003	0.508	< 0.001	0.706
pHKCl
r	−0.738	−0.673	−0.346	−0.823	−0.877	−0.872	−0.931	−0.916	−0.927	−0.788
P	< 0.001	0.001	0.135	< 0.001	< 0.001	< 0.001	< 0.001	< 0.001	< 0.001	< 0.001
% DM
r		0.193	0.115		−0.486	−0.507		−0.400	−0.715
P		0.415	0.629		0.030	0.022		0.080	< 0.001
Notes: ^aTwo-tailed tests.

Despite its clear relationship with pH, Al saturation cannot have been wholly determined by this parameter because (1) under all treatments except CS, Al saturation peaked together with pH in November 2002; and (2) during 2002, the only treatment with which pH differed significantly from that of controls at any time was BL, whereas it was other treatments that were associated with Al saturations differing significantly from those of control plots during this time. According to Haynes and Mokolobate,19 exchangeable Al levels can fall not only because of higher pH, but also because of complexation with soluble humic compounds and aliphatic acids released by decomposing organic waste. Such factors may explain why, for a given sampling date in Table 2, the difference between the greatest and least Al saturation was generally very considerable (20–40%), whereas the corresponding difference in pH was of relatively little practical relevance (generally <0.5 units in KCl or 1 unit in water), a contrast that is in keeping with Conyers et al.,5 who found that acid soil fertility is better predicted by Al-based measures than by pH.

Ca²⁺ always occupies more negative charges of CEC_e than any other cation except Al³⁺; it usually contributed more than Al³⁺ under treatments M and DS; and it always contributed more than Al³⁺ under treatment BL, which was also the only treatment under which Al³⁺ did not always contribute more than any other cation except Ca²⁺. The relative contributions of the other cations depended on treatment and date: that of Mg²⁺ exceeded those of K⁺ and Na⁺ in more than half of the samples, whereas the contribution of K⁺ also exceeded that of Na⁺ in more than half of the samples.

The exchangeable cation other than Al³⁺ that was most affected by treatment was Ca²⁺, with mean-over-dates levels ranging from 1.67 cmol_c kg⁻¹ in control plots to 3.21 cmol_c kg⁻¹ in BL plots, followed by Mg²⁺ (0.24 [CS] to 0.55 [BL] cmol_c kg⁻¹), Na⁺ (0.24 [M] to 0.29 [DS] cmol_c kg⁻¹), and K⁺ (0.26 [M] to 0.31 [BL] cmol_c kg⁻¹) (Table 2). Ca²⁺ concentration was in fact greater in BL plots than under any other treatment on 7 of the 10 sampling dates, which is attributable to Ca supplement in the diet of the chickens producing the chicken manure.33 DS plots had the third-highest average Ca²⁺ and K⁺ levels, the second-highest Mg²⁺ levels, and the highest Na⁺ levels; however, because of the marked leaching of Na in this area of high annual rainfall, and in keeping with previous findings in this area, Na⁺ content was not increased by DS treatment as much as is usual with DS.38

Under all treatments, Ca²⁺ levels were between 0.5 and 1.3 cmol_c kg⁻¹ lower in March 2004 than in March 2002, and between 0.1 and 0.6cmol_c kg⁻¹ higher in November 2004 than in November 2002, whereas opposite trends were shown by Mg²⁺ and Na⁺, and K⁺ levels were lower throughout 2004 than in 2002 (the only exception being the November K⁺ levels of BL plots) (Table 2). However, these temporal patterns must of course be considered jointly with the observed general fall in production (Figure 2) and probable changes in losses through leaching, both of which are attributable largely to inter-annual differences in weather conditions. In general, interannual variations in exchangeable cation contents were smaller than intra-annual variations and might be expected to be controllable, if necessary, by means of minor adjustments in the amendment dose.

Figure 2. Forage production under each treatment for each year of the study. For each year, the yields of treatments labeled with different letters differ significantly (P < 0.05).

Table 4 lists annual averages of Ca/Mg, K/Mg, and N/K ratios under the various treatments. The Ca/Mg ratio showed no statistically significant variation in time. Overall, the best Ca/Mg ratios (optimal < 10) tended to be those of BL plots, which were on average near optimal in 2004; the ratios of CS plots were significantly larger than those attained under any other treatment (P ≤ 0.011), and those of M plots were also significantly higher than those of BL plots (P = 0.019). The K/Mg ratio (optimal < 1.5) was significantly larger in 2002 than in the following years (P ≤ 0.001), falling to average values of 0.5–0.9 in 2004; overall, this ratio was significantly higher in control and CS plots than under other treatments (P ≤ 0.002) and significantly higher under treatment M than under treatment BL (P = 0.021), which afforded the best values. In 2002, the N/K ratio (optimal = 1–2.5) was below recommended values under all treatments, but the significantly larger 2003 values (P ≤ 0.001) were in all cases within the recommended range; furthermore, this parameter continued to improve, being significantly larger overall in 2004 than in 2003 (P = 0.010). Although BL plots had the best N/K ratios in 2002, this treatment was associated with the least improvement, and overall the average N/K ratio was the same under BL as in control plots, 1.1. The greatest improvement was shown by DS plots (mean 0.7 in 2002, 1.7 in 2004), but treatments DS, CS, and M were overall associated with larger N/K ratios than the control and BL treatments (P < 0.05).

Table 4. Ca/Mg, K/Mg, and N/K soil ratios under each treatment in each year of the study (means ± standard deviations) together with recommended values and the results of analysis of variance

Variable	Factors (Treatment, Date)	2002	2003	2004	Recommended Value/Range
Ca/Mg
C	a	7.4 ± 4.5^a	7.3 ± 1.3^a	6.5 ± 1.2^a	5 (see ref 52)
M	a,c	7.6 ± 4.0^a	9.4 ± 3.3^a	7.3 ± 1.0^a	–
CS	b	14.4 ± 17.4^a	8.4 ± 1.5^a	8.4 ± 1.6^a	–
DS	a	4.9 ± 1.8^a	7.5 ± 1.9^a	7.1 ± 1.4^a	< 10 (see ref 53)
BL	a,d	6.1 ± 0.9^a	8.3 ± 2.9^a	4.9 ± 1.1^a	–
K/Mg
C	a	1.8 ± 0.6^a,b	1.1 ± 0.3^a	0.9 ± 0.3^b	< 1.5 (see ref 52)
M	b,c	1.4 ± 0.7^a	0.7 ± 0.2^a	0.6 ± 0.2^a	–
CS	a	2.6 ± 2.3^a	0.9 ± 0.2^a	0.8 ± 0.2^a	–
DS	b	1.2 ± 0.5^a	0.6 ± 0.3^a	0.6 ± 0.2^a	–
BL	b,d	0.7 ± 0.2^a	0.6 ± 0.3^a	0.5 ± 0.2^a	–
N/K
C	a	0.7 ± 0.3^a	1.2 ± 0.4^a,b	1.3 ± 0.4^b	< 1–2.5 (see ref 52)
M	b	0.7 ± 0.3^a	1.2 ± 0.4^a,b	1.6 ± 0.4^b	–
CS	b	0.7 ± 0.^a	1.4 ± 0.4^b	1.5 ± 0.4^b	–
DS	b	0.7 ± 0.3^a	1.5 ± 0.5^b	1.7 ± 0.5^b	–
BL	a	0.9 ± 0.3^a	1.1 ± 0.4^a	1.2 ± 0.5^a	–
Notes: Analysis of variance factors. For each parameter and date, data labeled with different letters differ significantly (P < 0.05).

Available soil P levels were generally low, and marked retention of P was indicated by high values of pH_NaF. 39 Time-averaged levels were increased by treatments DS (mean 8.95 mg kg⁻¹ vs. 6.04 mg kg⁻¹ in control plots), BL (7.91 mg kg⁻¹), and M (7.25 mg kg⁻¹), but not by CS (6.09 mg kg⁻¹), despite the CS having a P concentration almost as high as that of DS (Table 1) and a value for the C/P ratio that, like those of the other organic wastes, lay within the range ensuring rapid mobilization of P when applied to soil.40 According to Haynes and Mokolobate,19 the sorption of P by the soil after the application of P-rich organic waste is limited by competition with other decomposition products for binding sites and by the increase in surface negative charge associated with increased pH. The inability of treatment CS to increase available P in this study is accordingly attributable to its not having raised pH. However, under all treatments, P levels were lower throughout 2004 than in 2002 (the only exception being the March levels of BL plots) despite the lower production.

Summing up, nutrient contents for which between-treatment differences of practical significance were observed generally had their highest average values in BL plots and their lowest in control or CS plots, with a similar trend for pH and the opposite trend for Al saturation. Except for its marked effect on available P, treatment DS generally had much the same effect as treatment M. The relatively beneficial effect of treatment BL is in keeping with the results of previous studies of nutrient availability in soils treated with organic wastes.11,12,34,41–43 Hue and Licudine10 even reported a decrease in subsoil acidity after treatment with chicken manure and urban sewage sludge in a 14-day column leaching study.

Plant Production

As Figure 2 shows, production fell from 2002 levels of 6–12 t ha⁻¹ that are typical of Galician meadow to levels of 5–10 t ha⁻¹ in 2003 and 3–8 t ha⁻¹ in 2004. This decline is largely attributable to the weather (Figure 1): January 2003 had a mean temperature of 6 °C that was below the lower limit for meadow growth44 and was followed by a dry May and June that seriously reduced the size of the July crop. In 2004, an exceptionally dry winter was followed by high temperatures and low rainfall in June and July. However, the production of DS and BL treatments was less affected by the adverse conditions than that of other treatments: Although production by DS was significantly less than that of the other noncontrol treatments in 2002, together with BL production it was significantly greater than the others in 2003 and 2004. In the case of BL treatments this may have been partly because of lower Al saturation and correspondingly higher Ca²⁺, Mg²⁺, and K⁺ contents, but these factors do not account for the influence of DS treatment on production, and among all samples Al saturation was only significantly correlated with lower production in 2003 and July 2004 (Table 3). It is possible that DS and BL may have improved aggregate stability and water retention, as described by Haynes and Naidu.45 DS and CS treatments also supplied considerable quantities of water, but in the CS case this seems to have been offset by low Ca content (Table 1). The forage/crop production advantages of manure and water-treatment sludge over commercial fertilizers and liming agents have previously been reported by several authors.12,13,26,35,43,46

The crop initially established in these plots was a mixture of grass and clover; mixed pasture has agricultural and ecological advantages deriving from competition, complementarity, and nutritional interrelationships among species.47,48 Combinations of grasses and legumes, in particular, have greater nutritional value and a better year-round distribution of production than monocultures.49,50 In this study, treatment had no significant in fluence on the evolution of botanical diversity, which saw the virtual elimination of clover in favor of grass (Figure 3). This process is attributable to the neutralization of soil acidity and the supply of rapidly mineralizable N,51 to gether with the limitation of summer clover growth by retention of P. On the other hand, the combination of unusual dry weather and harvest low frequency could deplete the growth-competitive ability of clover.

Figure 3. Botanical composition of the forage produced under each treatment in each year. Treatments: C = control; M = mineral.

CONCLUSIONS

For these acid soils devoted to artificial meadow, granulated chicken manure outperformed mineral fertilizers in respect to all of the soil and production variables examined in this study. In particular, it significantly increased soil pH and, more markedly, Al saturation, thus reducing or eliminating the need for separate pH amendment. DS performed similarly to mineral fertilizer in most respects, but it led to greater available P levels and significantly greater forage yield. A mixture of weather conditions, joined with a diminution of clover competitive advantage by N addition, and lower P availability could explain the change in botanical composition, with a significantly less presence of clover.

ACKNOWLEDGMENTS

This work was supported by the Spanish Ministry of Science and Technology under projects AGF1999-0418-C02-02 and AGL200307385, by the Xunta de Galicia under project PGIDT01AGRO2E, and by a grant awarded to Mariana Matos-Moreira by the Fundação para a Ciência e a Tecnologia (Portugal).