Nutrient Cycling in Agroecosystems

, Volume 87, Issue 1, pp 137–149

Potassium uptake and requirement in organic grassland farming

Authors

    • Soil and Environment DivisionNorwegian Institute for Agricultural and Environmental Research
  • Sissel Hansen
    • Organic Food and Farming DivisionNorwegian Institute for Agricultural and Environmental Research
Research Article

DOI: 10.1007/s10705-009-9320-5

Cite this article as:
Øgaard, A.F. & Hansen, S. Nutr Cycl Agroecosyst (2010) 87: 137. doi:10.1007/s10705-009-9320-5

Abstract

Use of mineral fertilizers is restricted in organic farming. The aim of the present paper was therefore to study whether potassium (K) limits yields in Norwegian organic grasslands. The K status in soil and herbage on 26 organic farms was investigated, and the response to K application in six fertilization experiments was explored. Further, the relationship between soil K analyses and K release from soil was examined. K application to grassland on the investigated farms was generally low, giving negative field K balances on 23 of the farms. The soils were classified as low or intermediate in readily available K (KAL) on 23 of the farms. The mean K concentration for herbage samples from the first cut on these farms was 18.0 g K kg−1 dry matter. In fertilization experiments, K application increased the K concentration in herbage. However, there was no significant effect on yield, even when K concentration in herbage on plots without K application was low. The lack of significant yield response to K application can be explained by low amounts of crop-available nitrogen (N). There was a tendency for increased plant uptake from reserve K with increasing values of acid soluble K (K–HNO3) in soil. Separate K analyses of timothy (Phleum pratense) and red clover (Trifolium pratense) revealed that red clover showed better competitiveness for K than timothy in leys where N supply was limited.

Keywords

Available potassiumOrganic farmingPotassium applicationSoil analysisRed cloverTimothy

Introduction

In organic farming, potassium (K) fertilization is mainly based on recycling on-farm K by use of farmyard manure and compost. Usually some nutrients are imported to the farm in purchased fodder, but the amount of imported K often does not balance the K losses from the farm (Asdal and Bakken 1999; Askegaard and Eriksen 2000; Eltun et al. 2002; Watson et al. 2002; Korsaeth and Eltun 2008). Potassium is lost from the system by exported products and by internal losses on the farm from handling of fodder and manure, leaching from soil, etc. Negative K balances coincide with the fact that the soil K concentrations have been shown to decrease during conversion to organic farming (Løes and Øgaard 1997; Aasbø et al. 1999). The total K content in mineral soils is often high, usually between 0.2 and 3.3% of the soil dry matter (Scheffer and Schachtschabel 1998). The highest values are often found in clay soils in temperate regions, but depending on the mineralogy, a high K content may also be found in sandy soils. Release of K by chemical weathering varies widely among locations because of variations in soil mineralogy. Holmqvist et al. (2003) estimated weathering rates ranging from 3 to 82 kg K ha−1 per year in different Norwegian and Swedish soils. Obviously, the weathering rate is too low in some soils to balance a possible deficit in the K budget and there may be a need for extra K input.

The standards for organic farming set restrictions on purchased mineral potassium fertilizers, and their use needs dispensation from the organic certification body. Soil analyses and available K resources on the farm are important for evaluating the need for purchased K fertilizer. An earlier study on conventional grassland showed a fairly good correlation between the soil analyses routinely used to determine K in Norway (ammonium acetate lactate extractable K (KAL) and acid soluble K (K–HNO3 minus KAL)) and K release from soil (Øgaard et al. 2002). Experience has shown that KAL (comparable to exchangeable K) is often inadequate for assessing the need for K fertilizer. Several plant species may use significant amounts of K from sources that are not initially exchangeable (e.g. Sinclair 1979; Ghorayshi and Lotse 1986; Wulff et al. 1998). Release of K that is not initially exchangeable originates mainly from interlayer K (fixed K) in sheet silicates. Therefore, extraction with boiling 1 M HNO3, which extracts a part of the K in sheet silicates, is included as a routine method in Norway.

Legumes are important for nitrogen (N) supply in organic farming systems, and K is important for the ability of legumes to fix N (Duke and Collins 1985). The ability of clovers to compete with grasses for K in clover-grass mixtures is therefore essential for production in organic grasslands. Grasses and cereals have often been found to be better at exploiting K sources in soil than dicots (e.g. legumes) (Schön et al. 1976; Steffens and Mengel 1979; Kuhlmann and Wehrmann 1984), but to our knowledge this has not been investigated in N-restricted grassland.

The K relationships found for conventional grassland may not be valid for organic grassland, because of the usually restricted N supply on organically run farms (e.g. Berry et al. 2002), and the different botanical composition, with more clover in organic than conventional grassland. The N supply in organic farming mainly relies on farmyard manure, compost and N-fixing leguminous plants. Consequently, the yields are usually lower than on conventional grassland and the N concentration in herbage from organic grassland often is low (Ebbesvik 1998).

The aims of the present study were to investigate the:
  • K status in soil and the K and N concentrations in herbage in organic grasslands on mineral soils,

  • extent to which K limits yields in organic grasslands,

  • relationship between soil K analysis and K release from soil in organic grassland, and

  • ability of clover to compete with grasses for K in an organic grassland system.

Materials and methods

Two different data sets were used in the present study:
  • The first was based on analyses of a collection of soil and herbage samples from grassland on 26 organic farms with mineral soils located in four different regions of Norway (Fig. 1). The dataset was used to investigate the K status in soil and the K and N concentration in herbage and to calculate the field K balances. The farms had maintained organic plant production for more than 3 years.
    https://static-content.springer.com/image/art%3A10.1007%2Fs10705-009-9320-5/MediaObjects/10705_2009_9320_Fig1_HTML.gif
    Fig. 1

    Location of the farms in the farm study and the field experiments

  • The second data set was from six fertilization experiments in grassland on organic farms in southern and central Norway (Fig. 1). The response to K application, the relationship between soil K analysis and K release from soil and the ability of clover to compete with grasses for K in an organic grassland system were investigated.

Data collection from 26 organic farms

On each of the farms, three grassland fields were selected and sampled in 2001 and 2002. Yields were recorded and soil and herbage sampled on three subplots (6.0 m × 1.2 m) in each field. Subplots were parallel to each other with a maximum distance of 1 m between them. On most of the farms, the grassland was harvested twice each year, but on some farms there was only one harvest per year because they were in areas with a short growing season. For more detailed information about the study sites, see Govasmark et al. (2005). At the first harvest each year, the phenological stage of timothy was determined according to Moore et al. (1991) as mean stage by count of timothy, and varied from R1 (first spikelets visible) to R4 (anther emergence). The subplots received manure according to the fertilizer management of the individual farms. The amount of manure applied was taken from the farmers’ records. The K content in the manure was estimated from Norwegian standard values used in fertilizer planning (Bioforsk 2009). Soil sampling was performed after the first harvest in 2001. Ten soil cores from 0 to 20 cm depth were taken from each plot. Some soil characteristics are presented in Table 1. Samples of the crop were taken at each harvest and from each of the subplots. The samples from the three subplots were combined, giving one soil sample and one herbage sample from each field.
Table 1

Some soil characteristics of the sites in the farm study

Farm no.a

pH

C (%)

Clayb (%)

Siltc (%)

Sandd (%)

PAL (mg kg−1)

KALe (mg kg−1)

MgAL (mg kg−1)

CaAL (mg kg−1)

Acid soluble Kf (mg kg−1)

1

5.5

6.8

3

31

66

96

76

63

746

555

2

6.4

9.3

5

29

66

77

45

72

2,964

620

3

6.1

8.8

9

38

53

140

94

92

1,679

700

4

5.8

7.6

8

36

56

340

150

97

1,507

1,159

5

5.7

4.9

4

20

76

145

59

53

1,379

632

6

6.7

8.2

6

31

63

128

130

117

4,321

942

9

6.5

3.0

1

33

66

189

43

54

1,874

99

10

6.7

2.1

4

69

27

160

127

223

1,460

269

11

6.5

2.3

6

54

40

115

243

112

1,581

814

12

6.5

3.6

3

39

58

126

69

105

3,006

564

13

5.7

2.7

12

65

23

35

101

103

1,226

940

14

6.0

2.7

1

53

46

85

83

59

1,040

209

15

6.3

4.9

4

33

63

112

65

110

1,341

177

16

6.3

2.2

7

30

63

148

101

210

1,983

283

17

6.8

3.7

17

40

43

72

90

129

7,957

333

18

6.4

3.2

22

40

38

44

57

69

3,510

253

19

6.0

3.0

9

73

18

60

75

106

813

175

20

5.5

3.6

12

41

47

101

149

101

670

177

21

6.4

2.8

8

33

59

128

286

137

1,555

412

22

6.3

1.6

2

36

62

56

100

33

903

270

23

6.3

3.3

13

52

35

85

87

79

2,400

770

24

6.7

2.4

30

64

6

130

154

145

3,057

3,420

25

6.6

3.7

24

59

17

136

197

211

2,901

1,295

26

6.1

3.7

8

65

27

89

79

107

2,037

1,114

27

6.3

2.8

13

54

33

141

99

143

1,854

1,473

28

5.8

3.7

8

38

54

85

79

52

1,024

513

aFarm no. 7 was omitted from this study because the farmer did not provide the necessary records and farm no. 8 because it had organic soils

b<0.002 mm, 0.002–0.06 mm, d 0.06–2 mm

eIn fertilization planning, a KAL value of 0–65 is characterized as low, 66–155 as intermediate, 156–300 as high, and >300 as very high

fAn acid soluble K value <300 is characterized as low, 300–795 as intermediate, 796–1,200 as high, and >1,200 as very high

Fertilization experiments

Six fertilization experiments were carried out in 2004 in established grass-clover leys on organic farms. At four of the sites the treatments were repeated in 2005. The treatments were: 0 (K0) and 49 kg K ha−1 (K49) as potassium chloride (KCl). KCl was used as the K source because it is the only K source not containing other nutrients that may influence the yield. Except for the described K fertilization, the fields were not fertilized, except field no. 2, which received 20 tons of slurry ha−1 (0.19% K, giving 38 kg K ha−1) in 2004. The experimental design was a randomized block design with four replicates. The plots were 9 m × 2.5 m (harvested area 5.5 m × 1–1.5 m). Four of the fields were harvested twice each year, but field nos. 2 and 3 were harvested only once because they were in areas with a short growing season. The first harvest was taken approximately 1 week after the spikelets on timothy had fully emerged (growth stage R3, Moore et al. 1991). The dominant grasses were timothy (Phleum pratense), meadow fescue (Festuca pratensis) and perennial ryegrass (Lolium perenne). The clover consisted mainly of red and white clover (Trifolium pratense and T. repens, respectively). The relative contents of grass, clover, and dicotyledonous weeds were visually assessed shortly before harvest. After cutting, samples of total herbage were collected. In addition, on field nos. 1 and 2, separate samples of timothy and red clover were collected from the K0 treatment, and on field nos. 5 and 6 separate samples were collected from both the K0 and K49 treatments. Collection of separate samples of timothy and red clover is not complete regarding fields and treatments because of limited resources for the project. Soil samples from 0 to 20 cm depth were collected each spring before fertilizer was applied, and each autumn after the last harvest, 5–6 cores per plot. Soil and plant samples from two replicates were combined, giving two replicates for the soil and plant analyses. The samples were combined because it was necessary to reduce the costs of the analyses. Year of conversion to organic farming, age of the grassland field, botanical composition, and some soil characteristics are presented in Tables 2 and 3.
Table 2

The year of conversion to organic farming (the first year in which no conventional fertilizers were used), the age of the grassland field in 2004, and the botanical composition of first cut at the sites of the fertilization experiments

Field no.

Year of conversion

Age of grassland

Grass (%)

Clover (%)

Weeds (%)

2004

2005

2004

2005

2004

2005

1

1993

3

75

85

20

5

5

10

2

1994

1

80

75

20

20

<5

<5

3

1998

6

70

15

15

4

1984

1

60

30

10

5

1987

2

70

60

30

30

<5

10

6

1995

1

60

45

40

55

<5

<5

Table 3

Some soil characteristics of the sites of the fertilization experiments

Field no.

pH

C (%)

Claya (%)

Siltb (%)

Sandc (%)

PAL (mg kg−1)

KALd (mg kg−1)

MgAL (mg kg−1)

CaAL (mg kg−1)

Acid soluble Ke (mg kg−1)

1

6.1

3.6

14

69

17

47

94

207

1,061

421

2

6.3

1.7

2

22

76

101

59

84

801

132

3

6.1

3.0

6

50

44

54

35

53

1,015

285

4

5.6

1.5

5

41

54

112

61

54

366

381

5

5.8

4.8

9

26

65

127

72

76

1,218

1,097

6

6.4

2.4

15

33

52

37

64

233

1,796

232

a<0.002 mm, b 0.002–0.06 mm, 0.06–2 mm

dIn fertilization planning, a KAL value of 0–65 is characterized as low, 66–155 as intermediate, 156–300 as high, and >300 as very high

eAn acid soluble K value <300 is characterized as low, 300–795 as intermediate, 796–1,200 as high, and >1,200 as very high

In order to obtain more data from fields with acid soluble K above 1,000 mg K kg−1, yields and soil K were measured in organic grassland fields at three additional sites in 2006 with three plots of 10 m2 at each site. The plots were not fertilized. Harvesting and sampling were as described for the fertilization experiments. These data were only used in the regression analysis of the relationship between acid soluble K and K uptake from reserve K.

Plant and soil analyses

The soil samples were dried at 35°C and sieved through a 2 mm mesh. The samples of the crop were dried at 60°C and ground to a particle size of 1 mm.

The K concentration in herbage was determined by Inductively Coupled Plasma Atomic Emission Spectroscopy (ICP/AES) from a dissolution made in a microwave oven with ultrapure nitric acid and hydrogen peroxide in closed teflon vessels (Rodushkin et al. 1999). The N concentration in herbage samples from the farm study was determined by Kjeldal N, whereas in herbage samples from the fertilization experiments, the N concentration was determined by Near Infrared Spectroscopy (NIRS) (Fystro and Lunnan 2006).

The particle size distribution of the soil samples was determined by the pipette method (Elonen 1971). Soil pH was measured in a soil–water suspension (1:2.5 v/v). Total carbon (C) content was determined with a LECO CHN 1000 apparatus with an infrared (IR) detector (Nelson and Sommers 1982). In this study, total C was presumed to be primarily organic C, because the soils probably contained little carbonate. Easily available K, Ca and Mg (in the text abbreviated as KAL, CaAL and MgAL) were estimated by the method of Egnér et al. (1960), with the soil extracted by an ammonium acetate lactate solution (0.1 M ammonium lactate and 0.4 M acetic acid, pH 3.75) in the ratio of soil to solution of 1:20 (g:ml). The concentrations of K, Ca and Mg in the extracts were measured by ICP spectrometry. For comparison, KAL is often of the same magnitude as exchangeable K, extracted by 1 M ammonium acetate (pH 7.00). Acid soluble K was extracted by boiling the soil sample in 1 M HNO3 (ratio of soil to solution of 1:10 (g:ml) for 10 min (Pratt 1965, slightly modified). After filtration, the K concentration in the extract was measured by atomic absorption. As some of the extracted K is also soluble in the AL extraction, the KAL value was subtracted from the measured value. Acid soluble K is thus defined according to Eq. 1.
$$ {\text{Acid}}\,{\text{soluble}}\,{\text{K}} = {\text{K-}}{\text{HNO}}_{3} - {\text{K}}_{\text{AL}} $$
(1)

Calculations

In the fertilization experiments, K release from soil was calculated as the amount of K in harvested plant material minus applied K. Potassium released from KAL in topsoil (0–20 cm) was calculated from the difference between the KAL values in spring and autumn. The difference was converted to kg K ha−1 by using bulk densities calculated on the basis of the content of clay, silt, sand, gravel and organic matter (Riley 1996). For K uptake not originating from KAL in topsoil, the term reserve K is used in this paper (Eq. 2).
$$ {\text{Uptake}}\,{\text{from}}\,{\text{reserve}}\,{\text{K}} = {\text{K}}\,{\text{in}}\,{\text{harvested}}\,{\text{plant}}\,{\text{material}} - {\text{applied}}\,{\text{K}} - {\text{K}}\,{\text{released}}\,{\text{from}}\,{\text{K}}_{\text{AL}} \,\text{in} \,{\text{topsoil}} $$
(2)

Reserve K includes K uptake from subsoil in addition to K uptake from K not initially exchangeable in the topsoil. There also may be depletion of exchangeable K in subsoil, but this amount is probably small because of the low K application rates in organic farming. Therefore, it is assumed that the calculated uptake from reserve K is mainly attributed to K uptake from K fractions not initially exchangeable. The calculated K release from reserve K was dependent on the yield, and thus varies between years. Therefore, in the correlation analyses where uptake from reserve K was related to acid soluble K, average values of the two experimental years were used for the four sites with 2 years of data collection.

Statistical analysis

The effect of K application on yield and K concentration in herbage in the fertilization experiments was tested with analysis of variance, with treatment, year, harvest, field and replicate as class variables (GLM procedure, SAS Institute 1996). Nonsignificant interactions were removed from the model.

Simple linear regression analyses and correlation coefficients were calculated according to the least square method. The levels of significance are abbreviated *, ** and *** for 5, 1 and 0.1% levels of probability, respectively.

Results

Soil K values, field K balances and K and N concentrations in herbage in the farm study

The soils on 5 of the 26 organic farms sampled in 2001 and 2002 were classified as low in KAL, 18 as intermediate and 3 as high (Table 1). The values of acid soluble K are more dependent on soil mineralogy than earlier K fertilization. Acid soluble K values were classified as high or very high on 8 of the farms, and as low or intermediate on 18. The amount of applied K was low, and 38% of the recorded fields did not receive manure or other K sources. Only 15% of the fields received more than 100 kg K ha−1 in a year. Low amounts of applied K were reflected in negative field K balances. On average for the years 2001 and 2002, the K balance for the recorded fields was negative on 23 of the farms (Table 4). For all the samples from the first cut, the average value for the K concentration in the herbage was 18.0 (SD 4.5) g K kg−1 dry matter (DM), with 24% of the samples below 15.0 g K kg−1 DM. The average value for N concentration was 17.2 (SD 3.5) g N kg−1 DM, with 22% of the samples below 15.0 g N kg−1 DM. Average herbage K- and N concentrations for each of the farms are given in Table 4.
Table 4

Yield, K balance, K and N concentration in herbage from first and second cut and K:N ratio in herbage from first cut on each of the farms in the farm study

Farm no.

Yield (kg DM ha−1)

K first cut (g kg DM−1)

K second cut (g kg DM−1)

K balance (kg ha−1)

N first cut (g kg DM−1)

N second cut (g kg DM−1)

K:N first cut

1

216

13

14

41

22

27

0.6

2

301

14

14

−8

22

24

0.6

3

478

11

14

−37

15

22

0.7

4

598

12

19

−89

14

23

0.9

5

374

12

−44

14

0.9

6

573

21

25

−77

19

25

1.1

9

550

14

−35

15

0.9

10

508

18

24

−97

17

23

1.1

11

447

22

−66

16

1.3

12

557

18

−34

16

1.2

13

576

21

−107

18

1.1

14

542

16

−59

16

1.0

15

577

17

20

−106

17

23

1.0

16

441

22

26

−8

19

25

1.1

17

477

19

22

−97

21

34

0.9

18

472

16

18

−61

14

30

1.1

19

406

17

22

−37

17

28

1.0

20

408

22

25

2

20

26

1.1

21

448

24

31

−13

19

25

1.2

22

542

18

21

−45

14

17

1.2

23

583

19

23

−76

17

23

1.1

24

690

24

25

−88

16

20

1.5

25

594

23

25

−117

17

21

1.4

26

700

17

22

−37

16

22

1.1

27

758

23

26

−66

19

24

1.2

28

801

15

19

21

16

22

0.9

Average values for the years 2001 and 2002

K release from KAL in topsoil and reserve K in the fertilization experiments

The soils at the sites for the fertilization experiments were generally low in KAL; four of the sites had KAL values that are characterized as low, and at the two other sites, the KAL values were in the lower end of the intermediate class (Table 3). At three of the sites, the soils were also low in acid soluble K.

Without K application, total K uptake varied from 40 to 183 kg K ha−1 in the first year and from 58 to 107 kg K ha−1 in the second year (Fig. 2). The changes in the KAL values in the course of the growing season generally were small because of low initial values. Converting the change in the KAL values to kg K ha−1 in the upper 20 cm revealed that the K contribution from this fraction was below 40 kg K ha−1 for all sites (Fig. 2). There were large differences between the sites in K uptake from reserve K. Including data from three sites with acid soluble K above 1,000 mg K kg−1, there was a weak tendency for increased plant uptake from reserve K with increasing value of acid soluble K (R2 = 0.40, P = 0.069) (Fig. 3). However, as the figure shows, for acid soluble K values below 500 mg K kg−1 the variation in K uptake is large. Even with values below 300 mg K kg−1, which in fertilization planning is characterized as low, K uptake from reserve K amounting to about 100 kg K ha−1 was observed. With acid soluble K above 1,000 mg K kg−1, K uptake from reserve K was more than 100 kg K ha−1 in all cases, and varied between 110 and 143 kg K ha−1. The large variation in K uptake from reserve K was partly due to differences in the yield level, as increasing yield resulted in increased K uptake (R2 = 0.45**).
https://static-content.springer.com/image/art%3A10.1007%2Fs10705-009-9320-5/MediaObjects/10705_2009_9320_Fig2_HTML.gif
Fig. 2

Mean values, for each field, of K uptake from KAL and reserve K at zero K application

https://static-content.springer.com/image/art%3A10.1007%2Fs10705-009-9320-5/MediaObjects/10705_2009_9320_Fig3_HTML.gif
Fig. 3

Relationship between acid soluble K (K–HNO3 minus KAL) and K uptake from reserve K at zero K application in the fertilization experiments

Response to K fertilization in the fertilization experiments

The low K values in soil resulted in low K concentration in herbage with no K application (Table 5). The mean values for K concentration in herbage from the first and second cut were 13.3 (SD 3.6) and 13.1 (SD 4.2) g K kg−1 DM, respectively. Despite a low soil K status at most of the sites and a low K concentration in the crop, yields did not significantly respond to K application (P > 0.05, Table 6). However, K application resulted in a significant increase in the K concentration in herbage (P < 0.01, Table 5).
Table 5

Mean K concentrations in herbage from first and second cut at fertilization rates of 0 (K0) and 49 kg K ha−1 (K49) in each of two experimental years

Field no.

2004

2005

First cut

Second cut

First cut

Second cut

K0

K49

K0

K49

K0

K49

K0

K49

1

12.9 (±2.7)

15.5 (±0.5)

11.9 (±1.6)

12.9 (±0.8)

10.8 (±0.8)

16.3 (±0.9)

9.0 (±1.3)

11.1 (±1.8)

2

16.7 (±0.7)

16.7 (±0.3)

14.9 (±1.3)

18.2 (±0.2)

3

6.9 (±1.0)

10.0 (±0.4)

4

14.3 (±0.7)

14.5 (±0.3)

21.0 (±0.6)

23.0 (±1.2)

5

18.9 (±1.7)

20.9 (±1.5)

13.3 (±1.2)

14.3 (±2.7)

10.3 (±1.3)

17.1 (±1.3)

10.4 (±1.0)

12.9 (±0.6)

6

13.4 (±1.1)

17.1 (±1.3)

14.6 (±0.2)

18.7 (±1.8)

14.0 (±0.8)

19.5 (±0.2)

11.1 (±0.9)

14.4 (±0.0)

Field no. 2 and 3 were harvested only once. Field no. 3 and 4 were not run in 2005. Variations from the average value of two replicates in parenthesis

The values are expressed in g K kg−1 DM

Table 6

Mean herbage dry matter yields at fertilization rates of 0 (K0) and 49 kg K ha−1 (K49) in each of two experimental years

Field no.

2004

2005

K0

K49

K0

K49

1

7.38 (0.64)

7.06 (0.66)

5.62 (0.39)

5.89 (0.47)

2

4.84 (0.24)

4.98 (0.24)

5.11 (0.01)

5.27 (0.36)

3

5.73 (0.36)

5.99 (0.97)

4

7.48 (0.76)

6.87 (0.08)

5

11.1 (0.11)

11.2 (0.32)

9.94 (0.90)

9.77 (0.49)

6

7.48 (0.37)

7.99 (0.30)

8.44 (0.61)

9.58 (0.58)

Field no. 1, 4, 5 and 6 were harvested twice, and field no. 2 and 3 only once. Field no. 3 and 4 were not run in 2005. Standard deviations in parenthesis (four replicates)

The values are expressed in ton ha−1 year−1

Separate K analyses of timothy and red clover revealed that without K application, the K concentrations in the samples from the first cut were quite similar in timothy and red clover in all fields except field no. 2 in 2004 (Table 7). Potassium application increased the K concentration in both timothy and red clover (P < 0.001), but the increase was greater in red clover (P < 0.01). This was true for both years.
Table 7

Mean K and N concentrations in timothy and red clover from the first cut at fertilization rates of 0 (K0) and 49 kg K ha−1 (K49) in each of two experimental years

Year

Field no.

Potassium (K)

Nitrogen (N)

 

K0

K49

K0

K49

Timothy

Red clover

Timothy

Red clover

Timothy

Red clover

Timothy

Red clover

2004

1

11.8 (±2.7)

14.2 (±1.4)

n.d.

n.d.

12.3 (±1.4)

29.0 (±1.4)

n.d.

n.d.

2

11.2 (±1.7)

23.7 (±1.3)

n.d.

n.d.

8.8 (±0.3)

18.3 (±0.1)

n.d.

n.d.

5

18.0 (±1.7)

19.1 (±3.3)

20.8 (±0.6)

25.3 (±0.8)

12.8 (±1.0)

27.3 (±1.2)

11.3 (±0.0)

25.9 (±0.5)

6

16.0 (±0.6)

15.4 (±1.0)

16.6 (±0.0)

19.7 (±0.6)

11.3 (±1.1)

24.2 (±0.5)

11.0 (±0.7)

21.6 (±0.7)

2005

1

11.2 (±1.7)

12.3 (±2.7)

14.9 (±0.8)

17.9 (±1.4)

12.7 (±1.3)

29.3 (±0.5)

n.d.

n.d.

2

12.4 (±0.4)

11.5 (±1.5)

15.5 (±0.6)

18.4 (±0.4)

9.0 (±0.8)

18.7 (±1.2)

n.d.

n.d.

5

11.0 (±0.7)

11.7 (±2.4)

15.4 (±1.0)

16.8 (±0.3)

9.9 (±0.2)

23.7 (±2.3)

n.d.

n.d.

6

14.3 (±0.2)

11.0 (±1.5)

18.7 (±0.6)

17.7 (±0.1)

14.0 (±0.5)

27.9 (±0.6)

n.d.

n.d.

n.d. not determined

Field nos. 3 and 4 were not run in 2005. Variations from the average value of two replicates in parenthesis

The values are expressed in g kg−1 DM

Nitrogen concentration in herbage in the fertilization experiments

The N concentration in herbage also was generally low (Table 8). The average N concentrations for the first and second cuts were 13.7 (SD 3.1) and 18.4 (SD 2.9) g N kg−1 DM, respectively.
Table 8

Mean N concentrations in herbage from first and second cuts at fertilization rates of 0 (K0) and 49 kg K ha−1 (K49) in each of two experimental years

Field no.

2004

2005

First cut

Second cut

First cut

Second cut

K0

K49

K0

K49

K0

K49

K0

K49

1

15.1 (±1.4)

17.4 (±2.1)

15.8 (±0.7)

15.4 (±1.7)

13.0 (±1.8)

13.7 (±2.5)

15.7 (±1.1)

14.5 (±0.7)

2

14.3 (±0.6)

11.5 (±1.6)

13.0 (±1.4)

12.2 (±1.2)

3

12.1 (±0.4)

11.3 (±0.6)

4

10.0 (±0.4)

9.5 (±0.1)

18.6 (±1.0)

18.8 (±0.8)

5

14.4 (±0.7)

14.2 (±1.4)

16.6 (±0.4)

15.8 (±1.2)

12.5 (±1.5)

13.1 (±1.1)

19.4 (±1.3)

18.3 (±0.4)

6

13.7 (±1.5)

14.3 (±0.2)

23.2 (±0.7)

21.3 (±0.6)

20.4 (±0.6)

19.4 (±3.0)

21.5 (±0.0)

21.6 (±0.1)

Field nos. 2 and 3 were harvested only once. Field nos. 3 and 4 were not run in 2005. Variations from the average value of two replicates in parenthesis

The values are expressed in g N kg−1 DM

The separate analyses of timothy and red clover from the first cut showed that the N concentration was considerably higher in red clover than in timothy (Table 7). The average value in red clover herbage was 24.6 (SD 0.4) g N kg−1 DM, and in timothy 11.3 (SD 0.2) g N kg−1 DM. The N concentrations were not significantly influenced by K fertilization.

Discussion

K in soil and field K balances on organic farms

The K application to the recorded fields in the farm study was low. Consequently, the K balance was negative in the majority of the fields. The soils on 18 of the farms had KAL ≤ 100 mg K kg−1, that is, the KAL values were characterized as low or in the lower part of the intermediate class. The readily available soil K resources were therefore quite small on many of the farms. Earlier it has been found that KAL decreased during conversion to organic farming (Løes and Øgaard 1997; Aasbø et al. 1999). In addition to a reduction in the concentration of exchangeable K, it has been observed that cropping without K application increased depletion of K in the interlayers of the sheet silicates (von Boguslawski and van Lieres 1981; Hinsinger and Jaillard 1993; Møberg and Nielsen 1983). It is expected that with depletion of interlayer K, K release from soil will decrease with time. Further, Øgaard and Krogstad (2005) demonstrated that depletion of soil K resulted in increased K fixation capacity; that is, added K will be fixed in the interlayers of the sheet silicates and thereby be less available for plant growth. However, dependent on soil mineralogy, some deficit in the K balance is sustainable because of K release by weathering. Holmqvist et al. (2003) estimated weathering rates ranging from 3 to 82 kg K ha−1 year−1 in different Norwegian and Swedish soils.

Soil K analyses as predictors of K release from soil

In the fertilization experiments, K uptake from the KAL fraction was low because of low initial KAL values. There was a weak tendency for K uptake from reserve K to increase with increasing values of acid soluble K, but the variation for acid soluble K values below 500 mg K kg−1 was large. The relationship between soil K analyses and K uptake was partly masked by differences in growth potential. High yields reflect generally good growth conditions and thereby a better exploitation of the soil K resources. The growth potential varied among the different sites because of variations in climatic conditions, soil quality, and supplies of other nutrients, especially N. In conventional grass production with optimal supply of the important nutrients and thereby smaller variations in yield, a closer relationship was found between soil K analyses and K uptake (Øgaard et al. 2002). The large differences in growth potential and botanical composition in organic grassland make it difficult to predict K release from soil based on soil K analyses. Different amount of roots in the subsoil may also influence the relationship between acid soluble K and K uptake from reserve K, because the exploited soil volume thereby varies. Generally, few grass roots go deeper than 25 cm (Sveistrup and Haraldsen 1997), but clover roots go deeper, and the amount of clover may therefore influence K uptake from the subsoil. The four observations with acid soluble K values above 1,000 mg K kg−1 all released more than 110 kg K ha−1 from reserve K, thereby indicating that soils with a high value of acid soluble K have a high capacity to release K. These results are in accordance with the results from a study in conventional grassland, where it also was found that with acid soluble K values above 1,000 mg K kg−1, more than 100 kg K ha−1 were released from reserve K (Øgaard et al. 2002). However, as pointed out above, K depletion over time at such high levels is probably not sustainable. Andrist-Rangel et al. (2007) demonstrated that in addition to decreased concentration of exchangeable K, negative field K balance over time also resulted in some cases in decreased concentration of acid extractable K in soil. Further, decreased K concentration in ley herbage over time was also observed.

K and N in ley herbage and response to K application

There was no significant effect of K application on yield, despite a low K concentration (<15 g K kg−1 DM) in herbage from the plots without applied K. However, the K application resulted in a significant increase in the K concentration in herbage. Earlier, a Norwegian study on conventional grassland showed that the yield was increased by K application if the K concentration in the crop was lower than 15–18 g K kg−1 DM in the first cut and 17–20 g K kg−1 DM in the second cut (Lunnan and Øgaard 2005).

The lack of yield response in the present study was probably due to the restricted amount of crop-available N. The average N concentration in the herbage from the first cut was 13.7 g N kg−1 and thus far below the critical N concentration in forage grasses, which is often reported to be >20 g N kg−1 DM (Bélanger and Gastal 2000; Bergmann 1988). In a Danish organic farming system it was also found that K application without N application did not increase the barley/pea yield, despite a low K concentration in soil and in yield on plots without applied K (Askegaard and Eriksen 2002). As in the present study, K application only resulted in increased K concentration in yields. It is well known that the response to a particular nutrient depends on the level of other nutrients. Nitrogen and K are physiologically connected as K plays an important role in protein synthesis in plants (Blevins 1985). Full utilization of applied N therefore depends on an adequate K supply. On the other hand, the yield response to K supply depends on the level of N nutrition (MacLeod 1969; Mengel and Kirkby 1987; Fortune et al. 2004).

Herbage potassium and N concentrations decrease as plants mature. Therefore, the herbage K:N ratio has been used for evaluating K sufficiency in grass instead of herbage K concentration. No response to K application was found when the K:N ratio was above 0.8 in grass (Dampney 1992; Lunnan and Øgaard 2005). From Table 7 it can be calculated that the K:N ratio was above 0.8 in timothy samples from the first cut in both years for the fields where separate analyses of timothy and red clover were performed. The ratio varied from 0.9 to 1.4. The lack of significant response to K application in the present study, despite a low K concentration in the herbage, seems therefore to be explained by the low amount of available N for the crop, and thereby K:N ratios above 0.8 in the grass. The K:N ratio in red clover was much lower than in timothy because of N fixation in red clover and thereby a higher N concentration. The K:N ratio in total herbage from grass-clover ley is therefore influenced by the amount of clover and cannot be used as an indicator of K sufficiency when clover constitutes a significant part of the herbage.

In the farm study, the low K application was reflected in low herbage K concentrations in many cases. The average value for the samples from the first cut was 18.0 g K kg−1 DM, and 24% of the samples were below 15.0 g K kg−1 DM. However, the K:N ratios were below 0.8 only on three farms, despite that K and N analyses were performed on herbage samples containing clover (Table 4). This indicates that the growth in leys on organic farms often is not K limited even with low K values in herbage.

Potassium concentration in timothy and red clover

Comparable K concentrations in timothy and red clover and larger response to K fertilization by red clover than by timothy, as observed in the present study, are not in accordance with results from earlier studies. In an experiment where ryegrass and red clover were grown together, the K concentration was higher in the ryegrass and lower in the clover than when the two species were grown separately (Steffens and Mengel 1979). It has been claimed that in a mixed sward with grass and clover, low levels of available K may lead to the disappearance of the clover (Mengel and Kirkby 1987). The better competitiveness of grasses for K has been explained by the extensive root system of grasses in top soil. The discrepancy between earlier results and the present results could be explained by the restricted N supply in the present study. Because no N was applied in these fertilization experiments, the N2-fixing red clover had better access to N than did the grasses. This was demonstrated by the approximately double N concentration in red clover compared with timothy. As indicated by the higher N content in clovers than grasses, the biological N fixation was successful even in the fields with the lowest content of K. Different N supply may influence the competitiveness for K, such that red clover does not show lower ability than grasses to exploit available K when N supply from soil is restricted.

Conclusion

In the farm study, K supply to organic grassland was generally low, giving negative K balances in most cases. Further, the amount of available N was often suboptimal for growth. Restricted N supply influences the need for other nutrients, and even with low K concentration in herbage, K did not seem to restrict growth in the fertilization experiments. Clover, which is essential for production in organic leys, showed better competitiveness for K than timothy when N supply is limited. Therefore, considering only the effects on yield, the need for purchased K fertilizer to organic grassland therefore often cannot be justified. However, with significant negative K field budgets, extra K application is probably needed to maintain the soil’s long-term K-release capacity.

Acknowledgements

Financial support for this study was provided by The Research Council of Norway. Thanks to the Norwegian Extension Service and research stations in Bioforsk for performing the field experiments.

Copyright information

© Springer Science+Business Media B.V. 2009