ICEL. AGRIC. SCI. 27 (2014), 81-94

81

Soil phosphorus fractionation in Icelandic long-term grassland field experiments Thorsteinn Gudmundsson1) Sigurdur Thór Gudmundsson2) and Gudni Thorvaldsson1) 1)

Agricultural University of Iceland, Hvanneyri, 311 Borgarnes, Iceland [email protected] (Corresponding author), [email protected] 2)

Búgarður, Óseyri 2, Akureyri [email protected]

ABSTRACT Long-term fertiliser experiments on hayfields at Sámsstaðir, South Iceland, and Hvanneyri, West Iceland, provided the basis for investigations on phosphorus fractions, the fate of P fertilisers and P sorption in Icelandic soils. Total (Pt), inorganic (Pi), organic (Po), ammonium oxalate extractable (Pox), ammonium lactate extractable (PAL) and anion resin extractable (Pan) fractions were determined. The P sorption was measured and P sorption maximum (Smax) calculated. Ammonium oxalate extractable Si, Al and Fe were measured and the degree of P saturation (DPS) was calculated. We found all surplus applied P in the top 10 cm of the soil with the highest increase in the top 5 cm. While there was only a slight increase in Po, most of the surplus P was inorganically bound, with a strong correlation between Pt, Pi and Pox. Phosphorus saturation (Smax) was highly correlated with oxalate extractable Si and Al in the dry Silandic Andosol. However, in the Histic Andosol it was only highly correlated with Feox and not to Alox, indicating a different behaviour relating to redox conditions. Available P (PAL) increased with increasing P application mainly in the top 5 cm, but was not detectable at 10-20 cm depth. There was a good correlation between PAL and the degree of phosphorus saturation which only reached critical level in the top 5 cm with the highest P application of 39 kg ha-1 year-1. Water soluble Pan was substantially higher than PAL indicating that the phosphorus that had accumulated in the soil could be released and may be a useful source of P in future. Keywords: Long-term experiments, P fractionation, P balance, P sorption, DPS

YFIRLIT

Binding fosfórs í jarðvegi langtímatilrauna á túni Jarðvegur úr langtímatilraunum á túnum á Sámsstöðum og á Hvanneyri var notaður til að kanna bindingu og afdrif áborins fosfórs í íslenskum jarðvegi. Heildarmagn fosfórs (Pt) í jarðveginum var mælt og skipting fosfórsins í ólífræn sambönd (Pi) og lífræn (Po). Einnig var mælt hversu mikið losnar af P í ammóníum oxalati (Pox), ammóníum laktati (PAL) og með anjóna resin (Pan). Binding fosfórs var greind og hámarks aðsog (Smax) reiknað með Langmuir líkingunni. Si, Al og Fe voru mæld í ammóníum oxalati og mettunarstig fosfórs (DPS) reiknað.Við fundum allan umframáborinn fosfór í efstu 10 cm jarðvegsins en mest var þó bundið í efstu 5 cm. Það varð einungis lítil aukning á lífrænum fosfór á tilraunatímanum, fosfórinn var aðallega bundinn í ólífrænum samböndum og það er góð fylgni milli Pt, Pi og Pox. Milli mettunarstigs fosfórs (Smax) og Siox og Alox er góð fylgni í þurrum jarðvegi Silandic Andosol en í rökum Histic Andosol jarðvegi er einungis góð fylgni við Feox en engin við Alox sem bendir til að oxunar og afoxunarstig hefur veruleg áhrif á það hvernig fosfórinn

82

ICELANDIC AGRICULTURAL SCIENCES

binst. Nýtanlegur fosfór (PAL) eykst með auknum P-áburði, aðallega í efstu 5 cm jarðvegsins en í 10-20 cm dýpt er engin aukning merkjanleg. Það var góð fylgni milli PAL og mettunarstigs en það var einungis í efstu 5 cm þar sem hæsti áburðarskammturinn 39 kg P ha-1 var notaður sem fosfórinn var við það að ná fullri mettun. Vatnsleysanlegur fosfór, Pan, var verulega meiri en PAL sem gefur til kynna að fosfór sem safnast hefur fyrir geti losnað og nýst og verið mikilvægur fyrir framtíðina.

INTRODUCTION In Iceland the main soil types in agricultural use are deep Silandic and Gleyic Andosols and Histosols (Arnalds & Oskarsson 2009). Andosols are known to adsorb phosphorus (P) and require more P fertilisation than other soils. The surplus P accumulates close to the surface, which may be a potential risk because of surface runoff or if this layer is lost through erosion upon tillage (Auxtero et al. 2008, Parfitt 2009, Mejias et al. 2013). Phosphorus has been applied to Icelandic agricultural soil in surplus quantities for decades and accumulates in the top few centimetres in permanent hayfields (Gudmundsson & Sigvaldason 2000, Gudmundsson et al. 2005). In Andosols P is adsorbed by aluminium (Al) and iron (Fe) complexes (ligands), by allophane and by organic compounds (Dahlgren et al. 2004, Madeira et al. 2007, Parfitt 2009, Oburger et al. 2011). In fresh volcanic material appreciable amounts of P may be available due to rapid weathering of the primary P bearing mineral apatite (Dahlgren et al 2004). As the Andosols develop active Al, Fe and the adsorption capacity of the soil tend to increase. Ammonium oxalate is used to extract poorly ordered Fe and Al oxides and hydroxides and it also extracts P bound by Fe and Al minerals and complexes which often are the main minerals responsible for P fixation. Ammonium oxalate extractable P has been used to calculate the degree of P saturation and to estimate the soil P fixing capacity of Andosols (Beauchemin & Simard 1999, Pinheiro et al. 2007, Mejías 2013). Phosphorus fractions are frequently determined by sequential extractions where the method of Hedley et al. (1982) has been widely used and reviewed (Cross & Schlesinger 1995, Negassa & Leinweber 2009). Although

the Hedley fractionation is commonly used, this does not apply to Andosols. Negassa & Leinweber (2009) list over 100 studies on sequential P fractionation from different parts of the world and dealing with many different soil types. However, none of these studies were on Andosols. Otani & Ae (1999) used the Truog, Bray 2, Olsen and Citrate methods to estimate available P in Andosol and determined inorganically and organically bound P in the extracts. Pools of total, inorganic, ammonium oxalate extractable P and different methods for soil testing have been used to investigate the fate of P fertilisers, P retention characteristics of the soils, bio-available P and the soil phosphorus saturation degree (Baeauchemin & Simard 1999, Takahashi & Anwar 2007, Auxtero et al. 2008, Mejías et al. 2013). In the only study of P fractions in Icelandic soils so far, Helgason (2002) determined total, inorganic and organic P (Pt, Pi and Po) using the method of Saunders & Williams (1955) modified by Olsen & Sommers (1982). This method differs from the Hedley method in that it is not a sequential extraction and the organic and inorganic fractions are not further characterised. The P fractions are determined by extracting ignited and not ignited soil with sulphuric acid. Because of the different methods results cannot easily be compared. In Helgason’s (2002) study, the total P was in the range 440 to 2540 mg kg-1 in unfertilised areas and 1067 to 3129 mg kg-1 in fertilised fields. Organic P ranged from 0 to 1714 mg kg-1 and did not differ between fertilised fields and unfertilised areas. However Pi was lower in unfertilised soils with 207 to 1048 mg kg-1 compared with 996 to 1864 mg kg-1 in fertilised soils. This indicates that surplus fertiliser P is inorganically bound. This has also been reported for many other soils (Negassa &

P FRACTIONS IN ICELAND LT EXPERIMENTS

Leinweber 2009, Pätzold et al. 2013) and specifically for Andosols (Otari & Ae 1999, Takahashi & Anwar 2007). In general, Po was higher in West Iceland in the Helgason (2002) study, which is outside the country’s active volcanic area. The method of Saunders & Williams (1955) for total P fails to dissolve P firmly held in primary silicate minerals. Gudmundsson et al. (2005) compared total P determined by melting the soil with lithium borate dissolved in HNO3 and P dissolved in HCl in ignited samples for the “weatherable” fraction in a Gleyic Andosol. The “weatherable” fraction was 77 to 94% of the true total P without any obvious relation to depth or fertilisation. According to Dahlgren et al. (2004) the most commonly used tests to estimate available P in Andosols are the Truog method, most commonly used in Japan, followed by Bray 2, Mehlich 3, Olsen and anion exchange resin methods, used in the USA and many other countries. Investigating European volcanic soils, Madeira et al. (2007) compared extractions of P with H2O, CaCl2, Mehlich 3, Olsen, Bray2 and Egner-Riehm methods and found large differences in extractable P between methods. In Iceland, soil tests for agriculture use ammonium lactate extraction, the ALmethod of Egner & Rhiem (1960). For many years the Olsen method (Olsen et al 1954) was also used and Pálmason & Helgason (1990) compared P extracted by the Olsen and the AL-methods of Egner and Rhiem. They found that the AL-method extracted 2.53 (SE 0.18) and 1.78 (SE 0.18) times more P than the Olsen method in peat soils and freely drained soils, respectively. The coefficients of determination (R2) of linear regression between the two methods were 0.63 for the peat soils and 0.26 for the Andosols. Madeira et al. (2007) looked at the relationship between all extracts they tested and obtained a coefficient of correlation (R) of 0.74 by comparing the Olsen and the Egner Riehm methods for European Andosols, which was slightly better than Pálmason & Helgason found for Icelandic Andosols.

83

Iceland has large areas of unfertilised soils where phosphorus deficiency is apparent (Óskarsson & Brynjólfsson 2000) whereas hayfields have often received surplus P through fertilisation for many decades. The aim of this study was to investigate P fractions in fertilised hayfields and furthermore to investigate the fate of surplus P and to establish if it is organically or inorganically bound. Using the results we set out to see if there is a correlation between the various P fractions and the pedogenic oxides. Our final aim was to estimate the P sorption capacity and the degree of P saturation of the soils. MATERIALS AND METHODS Soil samples from three long-term fertiliser experiments on permanent hayfields were used. Two were at Sámsstaðir in South Iceland and one at Hvanneyri in West Iceland (Table 1). The oldest experiment started in 1938 (Sámsstaðir 1-49, Silandic Andosol) one in 1950 (Sámsstaðir 9-50, Histic Andosol) and one in 1970 (Hvanneyri 299-70, Hemic Histosol). The experiments at Sámsstaðir were stopped at the time of sampling; the one in Hvanneyri is still ongoing. In addition samples were collected just outside the experiments at Sámsstaðir as a reference for soils without fertilisation. The experimental areas were usually harvested twice a year, yield weight determined and samples taken for analysis. The grass was then removed from the site, including the unfertilised area. The grass samples were dried at 70° C and finely ground for analysis. Samples were digested by boiling in concentrated HNO3 and P determined using molybdenum yellow at 440 nm. Soil samples were collected from each plot using a 25 cm long conical core sampler, 3.1 cm in diameter. The cores were divided into 0-5, 5-10 and 10-20 cm segments at Sámsstaðir and to 0-5 and 5-15 cm depths at Hvanneyri. From each plot 15 to 20 cores were collected. The samples were mixed, dried at 35 °C and sieved through a 2 mm sieve. Subsamples for the C and N analysis, H2SO4 and NH4 oxalate

84

ICELANDIC AGRICULTURAL SCIENCES

Table soil type, type,fertilisers fertilisersapplied appliedannually, annually, sam- depths Williams (1955) Table1.1. Location, Location, soil sampling and year of as described by Olsen pling depths and year of sampling. 1-49 was sampled in 2005, & Sommers This involves sampling. 1-49 was sampled in 2005, duration 67years, 4 replicates; 9-50 was sampled (1982). in duration 67 years, 4 replicates; 9-50 was sampled in 2004, dura- weighing 1.0 g4 of soil into a crucible, 2004, duration 54 years, 4 replicates. 299-70 was sampled in 2006, duration 36 years, tion 54 years, 4 replicates. 299-70 was sampled in 2006, dura- ignite at 550°C for 2 hours, transfer to replicates. Soil types according to WRB 2006. tion 36 years, 4 replicates. Soil types according to WRB 2006.

a polyethylene bottle, add 50 ml 0.5 M H2SO4 and shake for 16 hours at room temperature. Parallel to this the same nf1) amount of soil was weighed into a 0P38 polyethylene bottle and shaken in the 0P49 26P same amount of sulphuric acid. The Sámsstaðir 9-50 Histic Andosol suspensions were sieved through Wat0 0 0 0-5 5-10 10-20 nf1) man No 42 filter paper and P in the 0P 70 0.0 74.7 0-5 5-10 10-20 13P 70 13.1 74.7 0-5 5-10 10-20 solutions determined using molyb22P 70 21.9 74.7 0-5 5-10 10-20 denum ascorbic acid blue at 880 nm. P 31P 70 30.6 74.7 0-5 5-10 10-20 Table soil applied annually, sampling depths year of 39P 1. Location, 70 39.3type, fertilisers 74.7 0-5 5-10 10-20 in the and ignited and not ignited sample is sampling. 1-49 Hvanneyri was sampled in 2005, duration 67years, 4 replicates; 9-50 was sampled 299-70 Hemic Histosol taken as Pt and in P , respectively. Then 2004, 299-70 sampled in 2006, duration 36 years, 4 i 0P duration 10054 years, 0.0 4 replicates. 100 0-5 was5-15 P = P – P . o t i 30P 100 types according 30 100 replicates. Soil to WRB0-5 2006. 5-15 Available phosphorus (PAL) was exnf = not fertilised area adjacent to the plots tracted using the ammonium lactate Treatment Annually applied kg ha-1 Sampling depths cm N ground P method after Egner et al. (1960) by shaking 5g extractions were in aK ceramic mortar 1-49 Silandic Andosol. of soil in 100 ml of 0.1 M ammonium lactate prior nfto1) analysis.0 Sámsstaðir 0 0 0-5 5-10 10-20 0.0 62.3 by using 0-5 5-10 10-200.4 M acetic acid for 2 hours and measur0P38 density70was determined and Bulk 100 70 0.0 62.3 0-5 5-10 10-20 49 ing P in the filtrate. cm3 0P cylinders for the 26P 70 26.2Histosol 62.3 in Hvanneyri 0-5 5-10 10-20 Water soluble P fixed by anion resin (Pan) and for the HisticSámsstaðir Andosol9-50 at Histic Sámsstaðir. For Andosol was 0 0 at Sámsstaðir 0 0-5 undis5-10 10-20extracted by the method of Sibbesen (1977 nf1) the Silandic Andosol 0P 70 0.0 74.7 0-5 5-10 &10-20 1978) as described by Burt (2004). In this turbed bulk density samples were not taken 13P 70 13.1 74.7 5-10 used10-20 Table 2. The P fractions and the methods of0-5extradition in this study. method 4 g of soil and 4 g of resin (DOWEX, 70 Instead 21.9 it was 74.7 measured 0-5 5-10 10-20 from22P each plot. by 31P 70 3 30.6 74.7 0-5 5-10 10-20 Matathon, type II, 510-610 µm spherical beds) weighing 200 cm air dried and sieved samples Fraction Extraction 39P 70 39.3 74.7 0-5 5-10 10-20 P 1) 0.5 M H2SO4 after ignition Pt = total in a Nitex nylon fabric bag with 300 µm pores and correcting to oven dry weight. Total C and Hvanneyri0.5299-70 Hemic Histosol M H2SO P = inorganic P 1) 4 were shaken in 100 ml of deionised water for 1) 100 NPio were determined by dry combustion using 0P 0.0 100 0-5 5-15 = organic P Pt - Pi 2) 30P oxalate P100 30ammonium 100 oxalate0-5 5-15 Pox = acid one hour followed by shaking the sample with Elementar vario MAX CN (Elementar AnP 3) toNH -lactate + acetic acid PALnf= =ammonium not fertilisedlactate area adjacent the 4plots another bag for 23 hours and the third bag for 4)GmbH). The pH was measured alysensysteme Pan = anion resin P deionised water with anion resin bags 24 hours. The P was released from the resin by using a and glass electrode inSommers a 1:21982. soil/water sus1) Saunders Williams 1954, Olsen and 2) Schwertmann 1964, 3) Egner & Rhiem 1960,a4)glass Sibbsen 1978, Burt 2004. shaking the bags in 80 ml of 0.5 M HCl for 30 pension using electrode. min after CO2 release had subsided. P was The methods of P fractionation used in this determined using molybdenum ascorbic acid study are outlined in Table 2. The determinablue at 880 nm. tion of total P (Pt), inorganic P (Pi) and organic The acid oxalate extraction (Pox) was adaptP (Po) followed the method of Saunders & ed from Schwertmann (1964) and Burt Table 2. 2. The PP fractions in this(2004) Table The fractionsand andthe themethods methodsof of extradition extradition used used in study. where 0.5 g of soil and 100 ml this study. of acid ammonium oxalate added were Fraction Extraction shaken in the dark for 4 hours, filtered 1) Pt = total P 0.5 M H2SO4 after ignition 
and Fe, Al, Si and Mn determined by Pi = inorganic P 1) 0.5 M H2SO4
Pt - Pi Po = organic P 1) ICP in Spectroflame D produced by 2) Pox = acid oxalate P ammonium oxalate Spectro, Germany. P was determined NH4-lactate + acetic acid PAL = ammonium lactate P 3) Pan = anion resin P 4) deionised water with anion resin bags using molybdenum ascorbic acid blue 1) Saunders and Williams 1954, Olsen and Sommers 1982. 2) Schwertmann 1964, at 880 nm. 3) Egner & Rhiem 1960, 4) Sibbsen 1978, Burt 2004. Treatment

1)

1)

Annually applied kg ha-1 Sampling depths cm N P K Sámsstaðir 1-49 Silandic Andosol. 0 0 0 0-5 5-10 10-20 70 0.0 62.3 0-5 5-10 10-20 70 0.0 62.3 0-5 5-10 10-20 70 26.2 62.3 0-5 5-10 10-20

P FRACTIONS IN ICELAND LT EXPERIMENTS

The determination of the P sorption isotherm is based on a method described by Graetz & Nair (2000) which is an adaptation of the method of Holford et al. (1974). One gram of soil was placed in a 250 ml polypropylene bottle with 50 ml of 0.01 M CaCl 2 containing a concentration of P added as KH2PO4 of 0, 50, 100, 150, 200 and 300 mg P L-1. Two to four drops of chloroform were added in each bottle to inhibit microbial activity. The bottles were shaken for 24 hours at room temperature. The soil suspension was filtered through Whatman no. 42 filter paper. P in the solution was determined by the molybdenum blue colour at 880 nm. The linear Langmuir equation was used to calculate the sorption maximum, Smax (Holford et al.1974). The Langmuir equation can be written as: C/S = (1/kSmax) + (C/Smax) (1) where S = S’ + S0, the total amount of P retained, S’ = P retained by the solid phase, S0 = P originally sorbet on the solid phase, mg kg-1 C = concentration of P after 24 h equilibration, mg L-1 S max = P sorption maximum, mg kg-1 k = Constant related to the bonding energy, L mg P-1

The degree of phosphorus saturation (DPS) was calculated as oxalate extractable P, Al and Fe using the formula of Schoumans (2000): DPS = Pox /0.5(Alox + Feox) x 100. where Pox , Alox and Feox are in mol kg-1.

(2)

85

The pools of P in soil were calculated using the amounts of P (mg kg-1) in all fractions and the bulk densities of each treatment and adding the pools in 0-5 and 5-10 cm at Sámsstaðir and in the top 15 cm at Hvanneyri. All analyses of variance were conducted with the ANOVA procedure in SAS (9.4; SAS Institute Inc., Cary, NC, USA). Two-way ANOVA were used on each of the P fractions, mg kg-1 (treatment + replicate), but one-way ANOVA for the total amount of P in the soil. The Tukey-Kramer test was used to test differences between each of the fertiliser treatments. Regression analyses were performed with the REG procedure in SAS. RESULTS The soils Some soil properties at the experimental sites are shown in Table 3.The Silandic Andosol (1-49) at Sámsstaðir had about 20 mg Alox kg-1. Total soil organic carbon (C) decreased from 109 to 66 mg kg-1 and bulk density increased from 0.54 to 0.75 kg L-1 at 0-5 and 10-20 cm depths, respectively. The pH was in the range of 5.5 to 6.0. The Histic Andosol (9-50) at Sámsstaðir overlies more than a 3 m thick deposit of organic and mineral material with 20 to 54% organic matter and a few distinct volcanic ash layers. The soil had a Gleyic colour pattern to the top and the organic material was fibrous. In

Table 3. Soil properties at the experimental sites. Values are means of all treatments, the

Table 3. Soil properties at thearea experimental sites. Values are means of all treatments, the untreated adjacent area untreated adjacent and replications. and replications. Depth cm 0-5 cm 5-10 cm 10-20 cm

0.54 0.66 0.75

N C C/N Siox g kg-1 g kg-1 g kg-1 Sámsstaðir 1-49 Silandic Andosol 5.5 109 8.3 13.2 10 5.7 85 7.0 12.1 11 6.0 66 5.7 11.6 13

0-5 cm 5-10 cm 10-20 cm

0.34 0.45 0.54

5.2 5.5 5.8

Sámsstaðir 9-50 Histic Andosol 144 10.8 13.4 7 84 7.1 12.0 15 64 5.4 12.1 17

15 22 23

41 74 51

1.02 0.64 0.70

0-5 cm 5-15 cm

0.31 0.33

4.5 4.4

Hvanneyri 299-70 Hemic Histosol 286 18.0 15.9 nd 226 15.2 17.4 nd

nd nd

nd nd

nd nd

nd = not determined

Bd kg L-1

pH H 2O

Alox g kg-1

Feox g kg-1

Mnox g kg-1

19 19 20

33 36 37

0.65 0.63 0.69

86

ICELANDIC AGRICULTURAL SCIENCES

the top 20 cm the C content decreased from 144 to 64 mg kg-1 with depth and the C/N ratio was 12 to 13 (Table 3), which is low for poorly

decomposed organic material but can be explained by the long-term N fertilisation. Corresponding to the decrease in organic C, the bulk

Table 4. Phosphorus fractions (mg kg-1), Smax (g kg-1) and DPS % molar ratio in the three long-term experiments. Total (Pt), inorganic (Pi), organic (Po), acid oxalate (Pox), ammonium oxalate (PAL) and anion exchange resin (Pan). The treatments were compared at each site, depth and extraction using Anova and the Tukey-Kramer test. Levels not connected by same letter were significantly different. (p≤0.05). Root MSE (RMSE) from the analysis are shown as well as phosphorus values from unfertilised area close to the experiments (NF). Phosphours fertiliser

Depth

Pt

Pi

Po

PAL

Pan

Smax

DPS %

0P38

0-5

Sámsstaðir 1-49. Duration 67 and 56 years. Silandic Andosol. 2360 a 747 a 1613 a 982 a 6a 32 a

0P49

0-5

1738 b

809 a

929 b

977 a

2a

26P RMSE

0-5

3839 c 216

2288 b 120

1552 a 181

2476 b 175

76 b 11

0P38

5-10

2925 a

907 a

2018 a

987 a

3a

0P49

5-10

1658 b

805 a

853 a

1064 a

4a

26P RMSE

5-10

3025 a 501

1538 c 138

1487 a 634

1924 b 109

29 b 6

0P38

10-20

2060 a

824 a

1235 a

1070 a

2

4.7 a

0P49

10-20

2027 a

729 a

1298 a

1045 a

7

4.6 a

26P RMSE

10-20

2200 a 434

1231 b 185

969 a 350

1277 a 132

9 6

6.2 b 0.5

9.6 a

4.8 a

7.8 a 0.5

12.6 B 0.6

16 a

10.7 a

4.7 a

118 a 31

9.0 a 1.4

5.0 a 235 b 31

4.9 a 9.0 b 0.4

NF1)

0-5

1902

960

942

1297

5

41

8.9

6.5

NF1)

5-10

1840

907

933

1245

8

29

8.7

6.5

NF1)

10-20

1671

827

845

1161

5

5.7

Sámsstaðir 9-50. Duration 54 years. Histic Andosol a 658 a 1287 a 1180 a 14 a ab 1569 a 2439 a 2012 a 39 ab b 2887 b 2250 a 3238 b 66 bc bc 3658 b 2284 a 4148 b 108 c c 5301 c 2203 a 6022 c 168 d 584 663 473 17

0P 13 P 22 P 31 P 39 P RMSE

0-5 0-5 0-5 0-5 0-5

1838 4008 5137 5942 7504 958

0P 13 P 22 P 31 P 39 P RMSE

5-10 5-10 5-10 5-10 5-10

1960 2725 2497 2740 3401 386

a a ab a b

905 1166 1441 1704 1872 277

a ab ab b b

1055 1559 1055 1035 1529 663

a a a a a

1176 1453 1193 1640 1976 269

a ab a ab b

9 13 13 18 32 16

a a a a a

0P 13 P 22 P 31 P 39 P RMSE

10-20 10-20 10-20 10-20 10-20

1883 1883 1614 2086 1990 256

a a a a a

822 938 1066 1185 1187 270

a a a a a

1061 945 707 899 945 195

a a a a a

1022 1180 973 1216 1184 168

a a a a a

10 12 13 13 13 5

a a a a a

NF1) NF NF1)

0-5 5-10 10-20

3440 1929 1227

0P 30 P RMSE

0-5 0-5

0P 30 P RMSE

5-15 5-15

1822 1165 969

1618 764 258

2557 1426 837

58 15 5

46 a

9.1 a

312 a

8.2 a

977 a 451

6.9 a 1.6

21 a

14.0 a

39 a

17.8 a

127 a 46

13.0 a 3.0

22 a

8.7 a

37 a

11.1 a

38 a 8

10.5 a 2.0

221 59 29

7.5 10.8 9.9

Hvanneyri 299-70. Duration 36 years. Hemic Histosol 1003 a 163 a 840 a 35 a 25 a 1808 b 1079 b 445 b 231 b 2887 b 108 118 71 19 70

8.5 a 10.9 a 1.4

965 a 1345 b 54

14.0 a 24.0 b 1.5

122 a 320 b 38

1) NF = not fertilised area adjacent to the experiment.

 

Pox

843 a 1025 b 66

11 a 29 b 4

11 a 41 b 4

5.9 11.4 14.9 21.8 27.2 3.9

a a ab bc c

3.7 4.6 3.0 5.3 6.1 1.3

a a a a a

3.8 4.4 3.2 4.6 4.4 0.6

ab ab a b ab

15.4 6.1 3.6

P FRACTIONS IN ICELAND LT EXPERIMENTS

density increased from 0.34 to 0.54 kg L-1 in the top 20 cm and pH increased from 5.2 to 5.8. The oxalate extractable Al, Si and Fe were higher than the minimum requirements for Andic properties and the soil was classified as an Andosol. The Feox maximum at 5-10 cm depth and Mnox maximum at 0-5 cm depth was an indication of low redox potential and segregation of Fe and Mn close to the surface. Bulk density and pH were slightly lower than in the Silandic Andosol. The Hemic Histosol (299-70) at Hvanneyri, West Iceland, had the lowest pH of 4.5 and 4.4 in the 0-5 and 0-15 cm layers and bulk densities of 0.31 and 0.33 at the same depths. The organic C was high, 286 and 226 g kg-1 at the 0-5 and 5-15 cm depths respectively and the C/N ratio of 15.9 and 17.4 distinguished the Histosol clearly from the Andosols (Table 3). There were no significant differences in C, N or acid oxalate extractable Si, Al or Mn between treatments. Only Feox showed high variability in 9-50, which was due to the variability in redox potential within the site and Fe segregation near the surface. The P fractions Total P in the unfertilised area (NF) and at 10-20 cm depth was an indication of the original or background P content (Table 4). In the Andosols Pt was mostly in the range 1700 to 2000 mg kg-1 and around 1000 mg kg-1 in the Histosol (Table 4). The main exception was the 0P38 fertilization treatment in the Silandic Andosol which had not received P fertilisation since 1938, but had a high P content in the top 10 cm. The Pt, Pi and Pox content at the 0-5 and 5-10 cm depths increased with increasing P application. However, this did not apply to the Po content, except in the Histosol. The increase in Pi and Pox with increasing P fertilisation and the positive correlation between Pt and Pi (R2 0.89) and Pox (R2 = 0.83) and a correlation between Pi and Pox (R2 = 0.87) (Table 5) indicated that most of the applied P remaining in the soil was inorganically bound. The Po values were in all cases higher in the fertilised

87

than in the 0P treatments at all depths, with the exception of 0P38, which had a much higher Po than other 0P treatments. The coefficient of correlation between Pt and Po (R2 = 0.62) also indicated that some, but significantly less, of the surplus P was bound in the organic matter than was inorganically bound. The Pox fraction was measured in the two Andosols and followed more or less the same pattern as Pi and was on the whole 14% higher than Pi. This indicated that Pox was not only extracting P from mineral surfaces but also a part of the Po or P in organic-mineral complexes. In the unfertilised area and in the 0P treatment, the PAL values ranged from 2 to 58 mg kg-1 in the top 5 cm but decreased to 2 to 11 mg kg-1 below the top 5 cm (Table 4). The main increase in PAL with increasing P fertilisation was in the top 5 cm and to a lesser extent at a depth of 5-10 cm. At 10-20 cm depth there was practically no change, which was consistent with no or only a small increase in other fractions (Table 4). Pan was 2-3 times higher than PAL, showing the ability of P to be extracted in water provided it is removed from the solution. The high coefficient of determination (R2 = 0.91) (Table 5) between Pox and PAL indicated that the soil test method of Egner et al. (1960) was extracting P that was bound in Fe and Al complexes. The correlation with Pan was weaker, possibly because of fewer samples analysed by the anion exchange resin. However, the high intercept at low PAL and Pan values showed that the inorganic P and oxalate extractable P was partially too strongly bound to be accessible in the weaker extracts. P retention (Smax) and P saturation degree (DPS) The P sorption maximum (Smax) calculated from the adsorption isotherms were in the range of 6.9 to 24 g kg-1 (Table 4). It was significantly higher in the 30P treatment than in the 0P treatment in the Histosol at Hvanneyri, but in the Andosols at Sámsstaðir there was no

88

ICELANDIC AGRICULTURAL SCIENCES

-1 Table fractions of DPS% with (R2 = 0.75) but for Table 5. 5. Regression Regression ofofP P fractions (mg(mg kg-1kg ) and) and of Smax (g kg-1)ofofall allplots plots and and depths that Fe were ox depths that were analysed. Si and Al the R2 was 0.35 analysed.

Equations Pt = 880 + 1.29 Pi Pt = 36.5 + 1.97 Po Pt = 792 + 1.18 Pox Pt = 2054 + 10.0 PAL Pt = 1941 + 4.53 Pan

R2 0.89 0.62 0.83 0.33 0.57

p