The results of an NPK-fertilisation trial of long-term crop rotation on carbonate-rich soil in Estonia

ABSTRACT

Soil is an element of crop cultivation that demands consistent fertilisation to compensate for the nutrients that are removed by the harvest. Changes in soil because of prolonged fertilisation can only be estimated by long-term field trials. Experiments in long-term field trial site Kuusiku (since 1965) include crop rotation of potato, late harvest barley, early harvest barley undersown with forage grasses (red clover + timothy), 1-year forage grasses, 2-year forage grasses, and winter rye. Various combinations of mineral and organic fertilisers were used to investigate the yield, soil humus, phosphorus, and potassium content (available and total) of the top- and subsoil. Fertilisation improved the yield of different crops by 1.3–2.6 times; meteorological conditions caused the yield to vary up to 6.4 times. The concentration of humus decreased 0.2% when not using inorganic and organic fertilisers; use of fertilisers increased the concentration of humus by 0.2–0.6%. The humus-rich subsoil (3.5% humus) contained less available phosphorus than humus-poor subsoil (humus 3.0%), which had 29 and 63 mg P_DL kg⁻¹, respectively. Grasses in crop rotation enriched the soil with organic matter and reduced the excess of nutrients remaining from previous fertilisation, thereby decreasing nutrient leakage and eutrophication of bodies of water.

KEYWORDS:

• Environmental safety
• humus
• leaching
• soil phosphorus
• phytoextraction
• yield

Introduction

Soil processes take place very slowly, and therefore, long-term trials are needed to investigate the influence of variable factors. Long-term fertilisation trials enable exploration of both the yields of agricultural crops and the influence of fertilisers on the soil properties, sustainability, and environment.

Long-term data about the influence of organic and mineral fertilisers on the crop yield, crop quality, soil quality, and environment in European geoclimatic conditions can be found in many papers (Rogasik and Schroetter 1999; Häusler and Hannolainen 2006; Albert and Grunert 2013; Kismányoky and Toth 2013; Merbach and Schulz 2013; Tajnšek et al. 2013; Körschens et al. 2014). Depending on agro-climatic conditions, these studies are focused on different crop rotations, such as corn–winter wheat–winter barley; sugar beet–spring barley–potato; or potato–early harvest barley undersown with red clover–clover–late harvest barley. The current long-term fertilisation study explains the impact of climatic conditions and long-term fertilisation on crop yields, soil and environmental safety during the transition from sea to continent. This study differs from the aforementioned studies in geoclimatic conditions, soil parameters and crop rotation, complementing the general understanding of changes in soil fertilisation. Long-term tests, the only way to monitor the vertical migration (leaching) of soil nutrients. In this study, attention has been paid to phosphorus and humus in different layers of the soil. Leaching of phosphorus is generally considered to be slight, less than 0.5% of the P added to the fertiliser (Kärblane 1975). If nutrients, including P, are less removed by the harvest (for example, due to unfavourable weather), and if the P-containing fertiliser is still given, then accumulates in the soil. Excessive phosphorus in the soil, however, puts pressure on ecosystems, causing eutrophication of water bodies, including the Baltic Sea (HELCOM 2005; Carpenter 2008). This study aims to explain the impact of long-term fertilisation on both the environment and soil sustainability.

Materials and methods

Study site and soil

Trials of crop rotation and fertilisation were established to determine the requirements for Estonian soil fertilisation. The current study is based on data from the long-term NPK fertilisation trial that was established by A. Piho (1973) in Kuusiku (Northern Estonia, geographic coordinates: longitude 58.584816, latitude 24.422128 and altitude 55 m) on sandy loam Calcisols (IUSS 2015).

Crop rotation consisted of six blocks as follows: potato (POT), late (late harvest) barley (LBAR), early (early harvest) barley undersown with red clover + timothy (EBARus), first-year forage grasses (1yFGR), second-year forage grasses (2yFGR), and winter rye (WR). The size of each trial plot is 56 m² (7.5 × 7.5 m).

The experiment has 21 fertiliser treatments (four NPK fertiliser and two FYM levels) and 6 crops, in two repetitions. From 1975 to 1992 there was the whole crop rotation in space and time and each year there was a total of 252 plots. Since 1993 there is the crop rotation in time only – each year there is a total of 84 plots, two agricultures per year. Rotation takes 3 years.

The following are test treatments:

Table

1. N0P0K0	7. N2P1K2	13. N2P3K2	19. N2P2K2 + FYM2
2. N1P0K0	8. N2P2K1	14. N3P3K3	20. N3P2K2+ FYM2
3. N1P1K0	9. N2P2K2	15. N4P3K3	21. N4P3K3+ FYM2
4. N1P0K1	10. N2P3K2	16. N2P2K2+ FYM1
5. N1P1K1	11. N3P2K2	17. N3P2K2+FYM1
6. N2P1K1	12. N3P2K3	18. N4P3K3+FYM1

The amount of fertilisers, depending on the needs of different crops, are given in Table 1.

Table 1. The fertilisation in a long-term NPK-fertilisation trial in 1975–2015 in the Kuusiku trial area.

Crop rotation	Fertilising: NPK, kg ha^−1; FYM^a, t ha^−1
	N1	N2	N3	N4	P1	P2	P3	K1	K2	K3	FYM1	FYM2
Potato	40	80	120	160	18	36	54	35	70	105	30	60
Late barley	35	70	105	140	9	18	27	25	50	75	0	0
E. barley (unders)	20	40	60	80	26	52	78	30	60	90	0	0
1y forage grasses	0	0	0	0	0	0	0	0	0	0	0	0
2y forage grasses	40	80	120	160	0	0	0	30	60	90	0	0
Winter rye	45	90	135	180	13	26	39	30	60	90	0	0

^aThe nutrient content of manure (FYM): N (nitrogen), 2.8 kg^–t; P (phosphorus), 0.6 kg^–t; K (potassium), 4.7 kg^–t.

Traditional agrotechnical measures were used: the soil was ploughed to a depth of 0.22 m, and chemical plant protection measures were also used. Mineral fertilisers were applied during soil tillage before sowing (amounts shown in Table 1); manure was applied by autumn ploughing for the following year’s potato. The following fertilisers were used: ammonium salpeter, superphosphate, potassium chloride, and solid farmyard manure (FYM). Cereal straw was returned to the soil by ploughing. Application of fertilisers, sowing of crops, and harvesting were performed at the optimal time.

Initially, the crop rotation included all the above-mentioned six crops. Since 1993, only two crops were rotated annually, so the whole crop rotation was completed in 3 years. Therefore, different crops were harvested in different years: potato, 21 y; late barley, 24 y; early barley, 22 y; first-year and second-year forage grasses, 22 y; and winter rye, 18 y.

The varieties used in the trial are well suited for the local climate; they are replaced with the new varieties from time to time. The population variety of winter rye ‘Vambo’ was used in trials for 40 years and was then replaced (2015) with the hybrid rye variety ‘Palazzo’.

Soil sampling and analysis

Soil samples were taken using a soil drill at each plot from the 0–0.2 m of topsoil and from the 0.2–0.4 m of subsoil after harvest; each sample consisted of 15 subsamples. Soil samples were taken at least once per crop rotation. The soil samples air-dried and were sieved to <2 mm. The initial status of soil properties was the following: humus content was determined according to Tjurin (Vorobyova 1998), 2.7% (organic matter carbon C_org × 1.72); pH_KCl, 6.6; available phosphorus (P_DL) from double lactate extraction (Method 1977), 14 mg kg⁻¹ (estimated as the low content); and available potassium (K_DL), 96 mg kg⁻¹ (estimated as medium content). P_DL and K_DL were determined using double lactate (DL) extraction, Al and Fe were determined using the Mehlich III extraction (Mehlich 1984); Soil humus from soil samples taken after potato harvesting. In addition to the periodic agrochemical testing, the following investigations were performed: soil total phosphorus (P_tot) and total potassium (K_tot) were determined according to the ISO 11466:1995 ‘Soil quality – Extraction of elements soluble in agua regia’. The determination was performed by the ICP-OES (Inductively Coupled Plasma Optical Emission Spectrometer).

Sampling and analysis of crop yields

To determine the catch and its dry matter and chemical composition (N, P, K), the harvest was harvested at 11.5 m² (2.3 × 5) taken for 1 kg of analysis. The yield and dry matter of the yield were determined for all crops. The average dry matter (DM) yield of each type of fertiliser is shown in Table 2. Only one variety of rye was grown, and thus, depending on yields from all fertilisation, treatments are represented by rye. In other cases, the exact crop variety differed over the years. The fertiliser balance was determined by the amount of nutrient added wiht the fertiliser – the amount removed with the harvest. For an objective average content of chemical elements for practical use in the context of changing weather conditions, the chemical composition (N, P, K) of the harvest was determined several times. The amount of nutrients (NPK) in harvest was determined during 6–9-year measurements (1981–1989) for the following fertilisation treatments: N₀P₀K₀, N₁P₁K₁, N₂P₂K₂, N₃P₃K₃, N₄P₃K₃, N₂P₂K₂+FYM1, and N₂P₂K₂+FYM2. P and K were determined according to the EVS-EN 15621:2012, and N to the EVS-EN ISO 20483:2013.

Table 2. The average crop yield depending on fertilisation (1975–2015, Kuusiku).

Treatment	Field crop and yield, t ha^−1 , DM
	POT^a	LBAR	EBAR (us)	1y FGR	2y FGR	WR
1. N0P0K0	17.2	2.0	1.4	5.4	3.1	1.9
2. N1P0K0	17.7	2.6	1.7	4.7	3.8	2.2
3. N1P1K0	22.5	2.8	2.2	5.4	5.1	3.3
4. N1P0K1	19.7	2.7	1.9	4.9	4.0	2.7
5. N1P1K1	24.0	3.0	2.5	5.8	5.2	3.3
6.N2P1K1	24.8	3.4	3.0	5.5	5.5	3.5
7. N2P1K2	26.9	3.7	3.0	5.6	5.6	3.8
8. N2P2K1	27.6	3.7	3.1	6.2	6.0	3.9
9. N2P2K2	26.6	3.5	3.2	5.9	5.9	3.9
10. N2P3K2	27.2	3.7	3.0	5.7	5.6	3.8
11. N3P2K2	26.5	3.7	3.3	5.7	5.8	3.8
12. N3P2K3	26.6	3.7	3.4	5.5	5,7	3.9
13. N3P3K2	29.4	3.7	3.4	5.6	5.6	3.7
14. N3P3K3	31.0	3.8	3.6	6.5	6.4	4.1
15. N4P3K3	31.2	3.8	3.8	5.6	5.8	3.7
16. N2P2K2 + FYM1	29.4	3.7	3.1	5.9	5.9	3.8
17. N3P2K2 + FYM1	31.9	3.9	3.6	6.3	6.4	4.0
18. N4P3K3 + FYM1	32.3	3.9	3.8	5.9	6.0	3.9
19. N2P2K2 + FYM2	30.8	3.8	3.3	6.0	5.9	3.9
20. N3P2K2 + FYM2	31.9	4.0	3.6	6.3	6.3	4.0
21. N4P3K3 + FYM2	31.6	3.7	3.8	5.7	6.0	3.8
LSD0.05	2.11	0.24	0.32	0.20	0.40	0.27

^aPOT – potato, LBAR – late harvest barley, EBARus – early harvest barley undersown with red clover + timothy, 1yFGR – first year forage grasses, 2yFGR – second year forage grasses, WR – winter rye

To determine the factors that influence crop yields, the following treatments from the long-term NPK fertilisation trial were selected: unfertilised (N₀P₀K₀) from the fertilised treatments; and the treatment that yielded the largest average crop during the last 40 years of the trial.

The current paper includes long-term research data from 1975–2015, excluding the humus content in soil. Data for soil humus content have already been collected starting in 1965.

Calculation of hydrothermic coefficient

For Estonia, the transitional climate ranges from marine to continental, which are characteristic throughout the country. The average annual temperature is 6.7°C and the average precipitation is 696 mm. The average water evaporation per year is less than the precipitation. Temperatures and precipitation were measured during the vegetation period (Table 3) using an automatic weather station near the test. The conditions for plant growth and development are more specifically characterised by a hydrothermal coefficient (HTK), also called conditional moisture balance, which takes into account the combined effect of temperature and precipitation. Plants stress is higher than high temperatures, no water or low temperature and plenty of water. HTK was calculated using the sum of precipitation (S) and the sum of air temperatures (T) for the same period using the following formula: HTK = (S/0.1) × T. The period is considered to be dry if the HTK is between 1.0 and 0.6 and very dry if the HTK is 0.5 or less. If HTK exceeds 2.0 (Kivi 1998; Keppart et al. 2009), the period is considered to be too wet for crop cultivation.

Table 3. Temperature, precipitation, and hydrothermic coefficients during different vegetation periods (1975–2015, Kuusiku).

Period, years	Months
	April	May	June	July	Aug	Sept	Oct
Temperature, °C
1974–1983	3.5	10.3	13.9	15.7	14.9	10.5	4.5
1984–1993	4.2	10.3	14.5	16.3	15.1	10.2	4.1
1994–2003	5.3	10.5	14.9	17.3	15.5	10.0	4.7
2004–2013	4.8	11.1	14.5	17.7	16.2	11.5	6.0
2014–2015	5.4	10.4	13.3	16.2	17.8	12.0	4.9
Normal	3.1	9.5	13.9	16.6	14.7	10.4	5.0
Precipitation, mm
1974–1983	42	38	68	81	68	85	67
1984–1993	29	41	60	99	99	97	56
1994–2003	47	49	128	122	109	70	74
2004–2013	32	52	72	81	110	69	89
2014–2015	37	45	75	115	47	55	19
Normal	37	47	66	83	90	81	70
Hydrothermic coefficient
1974–1983	3.9	1.2	1.6	1.7	1.5	2.7	4.8
1984–1993	1.6	1.5	1.5	2.4	2.3	3.5	4.4
1994–2003	2.9	1.5	2.4	2.3	2.3	2.3	5.0
2004–2013	2.2	1.5	1.6	1.5	2.2	2.0	4.8
2014–2015	2.3	1.4	2.3	2.3	0.9	1.5	1.3
Normal	4.0	1.6	1.6	1.6	2.0	2.6	4.5

Statistical analysis

The correlation and variance analyses were used, and the median, variation coefficient (VC), and the standard deviation (SD) were calculated. The likelihood of a difference between treatments was found at a level of 95% confidence in the LSD_0.05 (Least Significant Difference) assay.

Results and discussion

Effects of long-term fertilisation on crop yields

The crop yield depends on several factors (e.g. fertilisation, weather conditions, crop variety, soil conditions). Investigation of the effect of fertilisation and weather conditions on crop yield revealed that most of the arable crops (except EBARus) gave a maximum total yield during the long-term trial period with the N₃P₃K₃ level of mineral fertilisation. Nitrogen uptake was most effective in the case of undersown early spring barley and potato, in which case solid manure was applied in addition to the mineral fertilisers.

However, weather conditions had a bigger effect on yield than fertilisation (Table 4). Because of fertilisation, the yield increased 1.3–2.6 times, but because of weather conditions, the yields differed by 2.3–6.4 times, depending on year. Rye produced a relatively stable yield and was not negatively influenced by weather conditions; the most unstable yield that was strongly influenced by weather conditions was characteristic for late barley.

Table 4. The average crop yields depending on fertilisation and year (1975–2015, Kuusiku).

Field crop	Treatment	Yield, t ha^−1, DM
		Max–min	Median	SD	VC, %	Fertilisation^a	Year^b
1. Potato	Control	26.9–7.0	16.8	15.7	91.3	–	3.8
	N4P3K3 + FYM1	52.4–14.1	31.8	25.2	78.3	1.9	3.7
2. L. barley^c	Control	2.0–0.8	1.9	0.6	28.4	–	2.5
	N3P3K3	7.7–1.2	3.5	1.4	37.5	1.8	6.4
3. E. barley (unsown)	Control	2.0–0.5	1.5	0.4	28.3	–	4.0
	N4P3K3	6.1–2.0	3.4	0.9	24.2	2.3	3.1
4. 1y forage. grasses	Control	9.6–2.5	4.8	1.8	33.3	-	3.8
	N3P3K3	12.1–3.1	6.3	2.0	30.8	1.3	3.9
5. 2y forage. grasses	Control	6.1–1.3	2.4	1.5	46.7	–	4.7
	N3P3K3	9.1–3.8	6.3	1.5	22.6	2.6	2.4
6. Winter rye	Control	3.3–0.7	2.0	0.8	47.2	–	4.7
	N3P3K3	6.4–2.8	3.9	1.1	31.2	2.0	2.3

^aFactor indicating how many times the yield was multiplied on fertilised plots compared to control plots.

^bFactor indicating how many times the largest yield was multiplied compared to the smallest yield.

^cL. barley – late harvest barley, E. barley – early harvest barley.

The influence of weather conditions as the one significant factor for the yield formation was analysed through relationships of weather and yield data. The results revealed that the investigated crops were divided into two groups according to the weather conditions that were necessary for crop development and growth.

The first group consisted of cereals: late spring barley, early spring barley and rye; and the second group included forage grass and potato (Figure 1). These two groups differed in solar energy and water use. Earlier investigations suggested that the cooler, the cloudier, and the rainier is in June and in the beginning of July (the period for stem elongation to ear formation), the better the grain yield (Hoffmann et al. 1966; Haller 1969; Sukamägi 1971).

Figure 1. The average positive (‘+’) and negative (‘−’) hydrothermic coefficients (HTK) for arable crops during the vegetation period from 1975 to 2015.

Note: The period is considered to be dry if the HTK is between 1.0 and 0.6 and very dry if the HTK is 0.5 or less if HTK exceeds 2.0.

Analysing the yield ability of different crops and HTK, we can conclude that unfavourable weather conditions for cereals are dryness in May and June, and a rainy August. For cereals, in years with the aforementioned unfavourable conditions, 21–46% of the potential yield was lost. For forage grass and potato, the most unfavourable conditions were wet April and June and dry May and August, in which case 20–35% of potential yield was not realised. In his earlier investigations, Haller (1969) mentioned that the yield of arable crops depends on water, air, and nutrition regime during the germination period. The species with powerful and wide-spreading root systems, which can accumulate water and nutrients from the wider area, are more resistant to the weather conditions, resulting in a larger yield (Haller 1969; Merrill et al. 2002; Dempewolf et al. 2014).

The crop yield is significantly dependent on the variety and its properties (Figure 2). For example, the yield of hybrid rye significantly exceeds the yield of population rye. Comparing the influence of fertilisation on the yield of different types of rye, we can see that the yield of hybrid rye showed a stronger correlation to phosphorus and nitrogen fertilisers than the population rye: correlation coefficients were 0.86 and 0.77, and 0.75 and 0.72, respectively (p < .01). Better usage of plant nutrients enables better yield formation: reducing yield formation costs and lowering the risks of environmental damage are the important factors here (Fageria et al. 2008). Aside from nutrient uptake, resistance to the unfavourable environmental conditions has the most significant effect on yield.

Figure 2. The average yield of winter rye (population ‘Vambo’ and hybrid ‘Palazzo’) in long-term NPK fertilisation trial depending on fertilisation treatments from 1975 to 2015.

In the same year as the winter rye exchange (2015), the early spring barley variety was replaced with another population variety. The new variety exceeded the maximum yield of the old variety by 0.3 t ha⁻¹, but here, the growth conditions, especially the precipitation amount, varied. The temperatures during the growth period were close to normal and were similar for both varieties, but the precipitation amount was 1/3 less for the new variety (460 and 295 mm, respectively). One characteristic trait of the new variety was better drought resilience because of its larger root system. We can conclude that the new variety is more resistant to the unfavourable weather conditions and is, therefore, more suitable for growing in the current area.

Figures 3 and 4 give an overview of the nutrient balance of different crops in crop rotation. The influence of different crops on NPK-balance in crop rotation in case of unfertilised (N₀P₀K₀) and fertilised (N₃P₃K₃) soil. We can see that in case of the 1.y forage grasses that don't get the fertiliser, most of the nutrients are removed from the soil. The storage and recovery of both N, P and K are different. Nitrogen is absorbed by the plants from the soil reserves, i.e. soil organic matter. Nitrogen fertiliser efficiency is influenced by several factors: timeliness, correct dose, climatic conditions, etc. The small effect of N₄ fertiliser (V treatment) is probably caused by the fertiliser being placed into the soil in one only operation, resulting in higher losses. In the current study, the amounts of P used were significantly higher than those used in plants (47%). Unused P accumulated in the soil. As a result of long-term fertilisation, the P_DL content in the soil was ‘high’ (45–105 mg kg⁻¹). Patch was removed by the P balance with the treatment of the II fertiliser (N₁P₁K₁), where the soil P_DL content was ‘average’. Potassium fertilisers were almost completely used by plants (91%), and in most of the treatments, the K_DL content in the soil was ‘average’. Soil P_DL and K_DL levels were closely related (p < .001) and correlated with the number of corresponding elements that were removed with the yield, which corresponded to correlation coefficients of 0.934 and 0.921, respectively.

Figure 3. The average use of nutrients N, P, and K in crop rotation by plants and its effect on soil PK from 1975 to 2015. Fertiliser treatments: I – N₀P₀K₀; II – N₁P₁K₁; III – N₂P₂K₂; IV – N₃P₃K₃; V – N₄P₃K₃; VI – N₂P₂K₂ + FYM1; VII – N₂P₂K₂ + FYM2.

Figure 4. Impact of different cultures on average NPK balance in crop rotation without fertilisation (N₀P₀K₀) and fertilised (N₃P₃K₃) soil.

Special attention should be paid to the nitrogen and phosphorus content because they can cause the eutrophication of water bodies. The yield of agricultural crops that removed the nutrients from the soil can vary substantially and therefore the amounts of nutrients removed from the soil with the yield also varies widely. Data from the current trial show that the nitrogen applied to the soil during the crop rotation (600 kg ha⁻¹) was removed by the yield (326–1046 kg N ha⁻¹), including nitrogen accumulated in the soil by Rhizobium trifolii by growing legumes (clover). From phosphorus applied to the soil (215 kg P ha⁻¹), 60–190 kg P ha⁻¹ was removed and from potassium applied to the soil (440 kg K ha⁻¹), 250–770 kg K ha⁻¹ was removed. The amount of nutrient removed depends largely on the crop yield, which depends to a great extent on weather conditions. The great variability of nutrient amounts makes it difficult to offer good suggestions for fertilisation. To avoid the leaching of nutrients that are not used by agricultural plants because of unfavourable weather conditions, new and innovative solutions are required.

Phytoextraction is the removal of excessive phosphorus by grassland plants; it is considered to be one of the most effective methods to reduce the soil P-content. The grasses are able to use well the nutrients left in the soil from by the pre-culture residues. The harvest removes nutrients from the soil, thereby reducing the environmental risk caused by leaching. By growing the grass-clover mixture (L. perenne + T. repens), 34 kg P ha⁻¹ y⁻¹ was removed by the yield. If there is too much phosphorus in the soil, grassland plants can help to remove it (Van der Salm et al. 2009; Timmermans Bart and van Eekeren 2016).

Unlike other crops that are grown for forage grass, PK-fertilisers were applied to the undersown early spring barley, and therefore, the soil PK-content increased in the first year. However, the P and K content in the soil was reduced in the following year. Depending on the yield of first year forage grass, phosphorus applied to the soil with fertilisers was not fully removed by agricultural plants in the crop rotation, and therefore the soil P_DL increased. Grassland uptake of nutrients in the first and second year was inconsistent: in the first year, the grassland plants removed more phosphorus from soil; in the second year, the forage grass removed potassium. These facts can potentially be useful for regulating excess nutrients in the soil.

In K₃-fertilisation level, a similar amount of potassium was removed from the soil to which the fertilisers were applied. Nutrition from potassium differs from other nutrients. Its importance in plant growth is explained mostly by participation in many physiological functions such as improving the uptake of plant nutrients and water and activating the synthesis of protoplasm and saccharides. Plants constantly use large amounts of potassium, and therefore, they can easily remain in deficit soil. The maximum yield can be achieved when the applied amount of potassium is similar to the amount used by agricultural plants (Sirendi 1981).

Although grassland plants can remove a lot of nutrients from the soil, they can also enrich the soil with nitrogen-rich organic matter with the help of azotobacters; crop rotation that includes grassland plants is, therefore, environment-friendly and guarantees soil yield ability.

Effects of long-term fertilisation on soil characteristics

Long-term NPK and farmyard manure fertilisation and unfertilised plots were investigated in the current trial to clarify the influence of fertilisation on humus content, which is a highly significant parameter of soil quality. The soil humus content decreased in unfertilised soil and in soil that was treated only with nitrogen fertilisers by 0.1 and 0.2%, respectively (Table 5). Larger applications of mineral NPK fertilisers increased the soil humus content by 7%, as follows: when mineral fertilisers were used together with farmyard manure at 30 t ha⁻¹, the soil humus content increased by 5–7%; when using farmyard manure at 60 t ha⁻¹, the soil humus content increased approximately by 12%; and where only mineral fertilisers were used, the soil humus content increased by 0.2%.

Table 5. The dynamic of humus content (%) in long-term crop rotation NPK fertilisation trial in Kuusiku.

Treatments	Years					LSD0.05
	1965	1976	1988	1995	2013
N0P0K0	2.7	2.6	2.5	2.6	2.6	0.09
N1P0K0	2.7	2.3	2.5	2.5	2.5	0.17
N1P1K1	2.7	2.5	2.6	2.6	2.7	0.11
N2P2K2	2.7	2.9	2.9	2.9	2.9	0.11
N3P3K3	2.7	2.8	2.7	2.7	2.9	0.10
N2P2K2+FYM1	2.7	2.7	2.8	2.8	3.1	0.25
N2P2K2+FYM2	2.7	3.0	3.0	3.1	3.3	0.31
N4P3K3	2.7	2.8	2.8	2.9	3.0	0.17
N4P3K3+FYM1	2.7	2.9	2.9	3.0	3.2	0.27
N4P3K3+FYM2	2.7	2.9	2.9	3.1	3.3	0.33

This was primarily because of forage grasses in the crop rotation; cereal straw was also ploughed into the soil. Mineral fertilisers together with farmyard manure at 30 and 60 t ha⁻¹ in the six-field crop rotation also increased the soil humus content, by 0.1 and 0.2%, respectively. High yields give the soil more organic matter and, in this case, especially leguminous grasses. According to Astover et al. (2009), relatively similar results were achieved in South-Estonia in long-term (1989–2007) three-field crop rotation (potato–spring wheat–spring barley) on Podzols (IUSS 2015), where the cereal straw was removed from the field, the humus content decreased from 1.7–1.5 in unfertilised soil; the humus content increased from 1.7–2.0 in the treatment where farmyard manure was used once per crop rotation in amount of 40 t ha⁻¹. Thus, we can conclude that 2 years of forage grass in a six-field crop rotation together with NPK fertilisers have similar influence on the soil humus supply as 60 tons of farmyard manure per hectare.

The ratio of soil organic carbon and total nitrogen (C/N) in unfertilised soil was 8.6; using the mineral fertilisers, the C/N ratio was between 9.2–9.4 and, using mineral fertilisers together with farmyard manure, the C/N ratio stayed between 9.0 and 9.3 (LSD_0.05, 0.27).

In unfertilised soil, long-term removal of plant nutrients by yield did not significantly change the P_DL and K_DL content, while in fertilised soil, large differences occurred in the P_DL and K_DL content depending on crop yield (Figure 5). Wide ranges in soil nutrient content can be explained by big differences in crop yield: yield is greatly influenced by weather conditions, especially by the amount of precipitation. Plants growing on unfertilised soil can use the nutrients released from soil resources until the content of P and K in the soil drops below the critical level where nutrients are released more slowly.

Figure 5. Available P_DL and K_DL contents in soil following different treatments during the study period from 1975 to 2013.

Our trial data revealed that phosphorus that was not used by plants was accumulating in the soil, increasing both the P_DL and P_tot content. If in the unfertilised soil, the P_tot content was 400 mg kg⁻¹, then it was 600–700 mg kg⁻¹ in fertilised soil. Potassium resources in the soil are larger than phosphorous resources, and fertilisation and non-fertilisation do not significantly affect the soil’s K_tot content.

Both adsorption and absorption of phosphorus in the soil have a great impact on plant nutrition and on eutrofication of water bodies, which, in turn, directly threatens environmental safety. Thus, the long-term use of phosphorus fertilisers was investigated. P_DL was released by mineralisation of soil particles, and removal by crop yield during one cycle of crop rotation from unfertilised soil was an average of 40 kg P ha⁻¹, depending on the crop 5–12 kg P ha⁻¹y⁻¹.

Several factors influence phosphorus in the soil. It can be bound chemically or physico-chemically (Fink 1991). Phosphorus is bound by Al, Fe, Ca, and Mg, but in the current study, iron compounds bound to phosphorus were most commonly observed. A significant correlation was discovered by P_DL and P_tot with Fe-content (Table 6). The phosphorus content was the highest in the soil’s clay fraction, which was also had the richest Fe and Al content (Kask and Niine 1997).

Table 6. Correlation coefficients between humus, the contents of the aluminium, iron, and the phosphorus on Calcisol (1975–2015, Kuusiku).

Phosphorus	Chemical properties soil
	Humus	Al	Fe	Ca
Ptot^b	0.98^b	0.61	0.84^a	−0.03
PDL	0.92^b	0.78	0.94^b	−0.12

^aSignificance levels, *p < .05; **p < .01.

^bP_tot total phosphorus, P_DL – determined using double lactate (DL) extraction.

Organic matter can also influence P sorption by binding it and reducing leaching (Debicka et al. 2016). During the trial, the soil humus content was increasing because of organic matter accumulation (e.g. cereal straw, forage grass and farmyard manure). The soil P_DL and P_tot content resulting from P-fertilisation, where the added phosphorus was not entirely used by agricultural plants, was also increasing. The content of P_tot in the soil was highest (700 mg kg⁻¹) with the N₂P₂K₂+FYM2 fertilisation treatment and the content of P_DL in the soil was highest with the N₃P₃K₃ and N₂P₂K₂+FYM2 treatments.

To explain the influence of soil humus content on migration of P adsorbed into the humus, the P_DL content in the topsoil and subsoil (at a depth of 0.2–0.4 m) in different fertilisation treatments and their connection to the soil humus content were investigated. Although phosphorus leaching is considered to be generally quite small (less than 0.5% of P added to the fertilisers) (Kärblane 1975), in our investigation of the subsoil (Figure 6), significant differences in content of available P resulting from different fertilisation were detected.

Figure 6. The influence of fertilisation and humus content on migration of P_DL during the study period (1975–2015).

Only with larger applications of mineral fertilisers, at fertilisation levels 2 and 3, the subsoil contained P_DL approximately at the medium level according to the fertilisation gradation table. Additionally, the study revealed that by increasing the humus content in the topsoil, the P_DL content in the subsoil started to decrease significantly. If the humus content exceeded 3.25% in the topsoil, then the subsoil contained 50% less P_DL. Kõlli and Tamm (2012) found that deep and humus-rich subsoil can reduce the danger of fertiliser leaching. This means that increasing the soil humus content is very important for improving soil fertility and for environmental protection.

Problems with phosphorus fertilisation can cause eutrophication of water bodies. Farmers affect the P-problem as they sometimes use too much fertiliser; thus, phosphorus that is not used by agricultural plants reaches water bodies. The soil parent material and water regime (Kask and Niine 1997) play an important role in the processes involving phosphorus in soil. According to Kärblane (1975), the compounds released into the soil by destruction of its mineral component and mineralisation of organic matter, and compounds added to the soil by fertilisation, will be partially absorbed by soil particles, and some will remain in the soil solution. Phosphorus will be absorbed biologically, chemically, and physico-chemically, leaving only a small amount in the soil solution. Phosphorous that leached out from the soil solution formed less than 0.5% of phosphorus added by fertilisers. Thus, we can conclude that the phosphorus problem in water bodies cannot be solved just by reducing fertiliser use; however, controlling their use is justified. Further, activities that reduce and inhibit the leaching of phosphorus into water bodies are relevant. The applications of P-fertilisers in the current trial were more than plants could use. The applied phosphorus amounts enable a yield of at least 6 tons of cereals per hectare. Additionally, the overdoses of phosphorus allowed us to investigate its movement in the soil. For the future, we recommend using varieties that can produce larger yields, are more resistant to the changing weather conditions, and use current measures that support the plants for larger yield formation.

The outcomes of the current study revealed that the humus content in the topsoil had a negative correlation with the P_DL content in subsoil (0.2–0.4 m), from which we can conclude that in soils that contain >3% of humus, the migration of phosphorus is smaller, and eutrophication of water bodies is avoided.

For the future, we recommend using varieties that can produce larger yields that are more resistant to changing weather conditions, and using agricultural approaches that support more effective uptake of plant nutrients. Additional study of other changes in the soil parameters such as water-holding capacity and soil microbial biomass activity are also recommended.

Acknowledgments

First, I would like to acknowledge Dr Arnold Piho (1924–1978) who was the creator of current trial, and whose death was premature. I would also like to thank Madis Häusler, MSc, from the Kuusiku Field Testing Centre at the Agricultural Research Centre, who was responsible for trial management for many years. I would also like to thank all the people who helped me with this work. The experiment belongs to the Estonian Crop Research Institute.

Disclosure statement

No potential conflict of interest was reported by the author.

Notes on contributor

Valli Loide (Tuisk) was born on 21 January 1949. She graduated from the Estonian Agricultural Academy in 1974 as an agronomist in the field of soil science. After graduation, the Estonian Institute of Agricultural Sciences started to work in the field of agrochemistry. Now it is the Estonian Crop Research Institute. Her Doctoral Thesis is ‘The content of available magnesium of Estonian soils, its ratio to potassium and calcium and the effect on the yield of field crops’, 2002. Her research focuses on soil fertility, plant nutritional needs, fertilisation, liming, and environmental issues. She has participated in projects with Estonian Agricultural University, Tallinn University of Technology and companies. She has also participated in several foreign projects, the most important of which are the project MOEL (co-operation project on fertiliser consumption in Central and Eastern European countries, 2007–2012), and Baltic Sea Region project Baltic Slurry Acidification (2016–2018). She has published more than 80 scientific articles, including 2 books. She has membership in Estonian Society of Soil Science, NJF, and Estonian Academic Agricultural Society. Her hobbies are sports (athletics, basketball and volleyball, table tennis, active movement), travel, hobby photography, nature, and theatre. She is married, has a son and two grandchildren.

Additional information

Funding

This work was supported by Ministry of Rural Affairs.