Sulphur and potassium fertilisation to organic ley
Sikrere grunnlag for tilråding av svovel- og kaliumgjødsling til økologisk eng
Part of Mineral-SIP
Project-plan
Goal
To get more knowledge to judge the need for, element of risk with and possibilities for supply of S and K with organic ley-cultivation in order to improve the plant growth and feed quality of the herbage for ruminants
Sub-goals
- Evaluate the S-supply for organic leys some places in Norway with low S-depositions
- Test the malate/sulphate-ratio as a method to evaluate the S-supply in ley
- Adapt a Norwegian model for estimating K- release from soil to organic farming practise
- Investigate short term effects of S and K fertilisation on the plants content of Cu, Mg, Ca, P, N, Mo
- Evaluate different S-sources
Hypothesis
We will test the following hypothesis:
- The amount of S available to plants are so low in parts of Norway that there is a risk of S-deficiency in organic leys
- Enlarged S-supply will increase the concentration of crude protein in herbage from ley with low S-supply
- Enlarged S-supply will increase the N/S-ratio in clover, but decrease the N/S-ratio in grass
- Enlarged S-supply in leys with low S-supply will decrease the content of Mo in timothy and red clover, but not affect the content of Cu.
- The malate/sulphate-ratio can be used as a method to evaluate the S-supply in ley
- The potential K- release from soil in organic ley can be estimated from analysis of K-Al and K-HNO3 and soil texture
- S and K fertilisation might create new imbalances/disorders in the supply of minerals for plants and ruminants
- The plant availability of S are lower in organic S-sources with high than with low C/S-ratio
- The plant availability of S are higher in organic S-sources where S is bound with C-O-S than with C-S
- [SH1]The effect of gypsum on plant uptake of S in grass-clover sward is just as large as the effect of Na2SO4
[SH2]
Methods
1. Quantification of the effect of S and K application in organic leys through field trials.
Seven 2 year field trials with a randomized complete block design with four replicates were placed in existing leys that are managed organically. Uniform leys with less than 10% dicotyledonous weed were chosen in areas that reflect different soil conditions and levels of S and K. The soils had a broad variation in acid soluble K (K-HNO3 minus K-AL) varying from about 200 to 1500 mg K kg-1 (or more). The K-AL values should be about 50-80 mg K kg-1 above the soil’s minimum level for K-AL varies due to soil texture (K-ALmin = 0.94 * (silt%+clay%) + 13.08) {Øgaard, Krogstad, et al. 2002 3628 /id}. [SH3]Areas where sulphur deposition is above 1000 mg S per m2 and yr according to {Hole & Tørseth 2002 3660 /id}, is avoided. Fields with levels of P-AL< 50 mg P kg-1, Mg-AL< 40 mg Mg kg-1 and/or Ca-AL< 750 mg Ca kg-1 are avoided.
I Norway the following classes are used to classify the content of soil-nutrients (all in mg kg-1 dry soil)
Class / K-HNO3 / K-AL / Ca-AL / P-AL / Mg-AL1 / Low / <300 / 0-60 / <500 / 0-20 / 0-20
2 / Middle / 300-790 / 70-150 / 500-990 / 30-60 / 30-50
3 / High / 800-1190 / 160-300 / 1000-1990 / 70-150 / 60-90
4 / Very high / >1200 / >300 / >2000 / >150 / >100
Field placement and field data (will be filled in autumn 2004):
Field nr / Sted / Time since convertion / Age of ley 2004 / % Clover1 / Skien
2 / Inderøya / 1 year
3 / Nord-Østerdal
4 / Aunegrenda
5 / Surnadal
6 / Tingvoll / 2 year / middels
7 / Toten / 1 year / 30-40%
Field nr
Soil type / %silt / %clay / K-AL min / K-HNO3
1
2 / Sandy silt / 50 til 60 / 5 til 10
3
4
5
6
7 / moldrik morene
Fertilisation treatments:
- S0K0. No fertilisation
- S1K0a. 20 kg S ha-1 given in gypsum (CaSO4 x 2H2O)
- S1K0b. 20 kg S ha-1 given in natriumsulphate (Na2SO4 x ?H2O)
- S0K1. 0 kg S and 49 kg K ha-1 given in potassiumchloride (KCl)
- S1K1. 20 kg S and 49 kg K ha-1 given in potassiumsulphate (K2SO4 x 2?H2O)
- S2K2. 40 kg S and 49 kg K ha-1 (Na2SO4+ K2SO4 = treatment3+treatment5)
Soil samples
At start of the trials a composite soil sample (0-20 cm depth) was taken from each field to determine the content of S (extractable with 0.016 M KH2PO4, according to Zhao and McGrath (1994)), total C and total N (analysed on a LECO CHN 1000 with an IR Detector, Nelson and Sommers 1982 and soil texture (by pipette method Elonen 1971). From each plot a soil sample consisting of 5 cores was taken for determination of easily releasable phosphorous (P-AL), potassium (K-AL) magnesium (Mg-AL) and Calsium (Ca-AL) according to Égner 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) and analysed on an ICP-AES[SH4]. K-reserves (K-HNO3) was determined by a modified method of Pratt (1965) by boiling the soil sample for 10 minutes in 1M HNO3 and recorded by flame photometry. pH was determined in a 1:5 w/w water-soil suspension. The soil content of K-AL was thereafter registered each spring and autumn at treatment 1, 2, 3, 4 and 5. Soil samples from two replicates were combined, giving two replicates from the soil analysis.
Harvest and plant samples for determination of DM and total content of S, K, P, Mg, Ca, Cu
The fields were harvested twice each growing season. The first harvest was approximately when the spikelets on timothy had fully emerged (growth stage R1-R2, {MOORE, MOSER, et al. 1991 2811 /id}) and second harvest normally 7-8 weeks after 1. harvest.
The herbage was cut with a two-wheel reaper (sampling area 5.5 m x (width of knife) m).
A composite sample of ca 500g, and a sample of at least 50g from each of timothy (Phleum pratense) and red clover (Trifolium pratense) from each plot. All samples were dried at 60ºC and milled to 1 mm mesh size[SH5]. The dry matter content was determined in the composite sample from each plot. The content of S, K, P, Mg, Ca, Cu, Mo were determined in composite-, timothy- and red clover samples at treatment 1,2,3,4,5. Samples from two replicates of each treatment were combined. Samples from treatment 6 were stored for possible later analysis. Whether timothy and clover plants from treatment 2 (gypsum) or 3 (Na2SO4) are analysed, depends on the fertilisation effect of gypsum and natriumsulphate. If there is no difference in yield and content of total S in composite sample, treatment 2 is analysed. If there is a larger effect of natriumsulphate, treatment 3 is analysed.
All samples were analysed on near infrared spectroscopy (NIR) in order to determine the amount of samples for chemical analysis needed to calibrate NIR to get a standard error of prediction (SEP) of 6-10% of the mean value for S, K, P, Mg, Ca”. The samples for chemical analysis were dissolved in a microwaveoven withultrapure nitric acid and hydrogenperoxide inclosed teflon vessels ({Rodushkin, Ruth, et al. 1999 3768 /id}) and assessed on ICP/AES[SH6]. When the NIR database was expanded with needed local (actual) samples to reach acceptable prediction accuracy (full cross validation) the content of S, K, P, Mg, Ca in the whole batch of samples was finally estimated.
Botanical composition was visually assessed in gras, clover and dicotyledon weed as a percentage of total yield shortly before each harvest, according to Nesheim (1986). The percentage of clover was also estimated using NIR.
[SH7]
Possible effects of S and K-fertilisation on the supply of Mo, Cu, Mg, Ca, P and N for herbage and ruminants is judged from the content of these elements in total herbage with different S and K- fertilisation. This judgement will not be fully, because we avoid fields with low levels of Mg, Ca and P.
5.1.1. The S-supply for organic leys in different places in Norway
The effect of S-fertilisation on S-supply is judged from increase in yield (kg DM ha-1), the content of total S (g S kg-1 DM) and N (g N kg-1 DM) in total herbage, timothy and clover and apparent recovery of S compared with treatment S0K0N0. Apparent recovery is defined as the amount of S in above-ground plant material in treatments receiving no fertilisation subtracted from the amounts in S-fertilised treatments, expressed as percentages of the amounts applied. On fields with a shortage of K treatments 5 and 6 were used to judge effect of S-supply and they were evaluated towards treatment S0K1.
5.1.2. Testing the malate/sulphate ratio as a method to evaluate S-supply in organic ley.
Samples are taken at each harvest each growing season from young, but fully grown leaves from red clover and timothy at fertilisation treatment 1, 2,3,5,6 for analysis of malate, sulphate, nitrate and phosphate. Whether plants from treatment 2 (gypsum) or 3 (Na2SO4) are analysed, depends on the fertilisation effect of gypsum and natriumsulphate. If there is no difference in yield and content of total S in total herbage, treatment 2 is analysed. If there is a larger effect of natriumsulphate, treatment 3 is analysed.
At to fields, field 6 and 7, additional samples are taken previous first harvest, using the phenological stage of timothy as defined by {MOORE, MOSER, et al. 1991 2811 /id} as criterion for sampling time. First sampling (vegetative leaf development, 4-5 leaf developed= V4-V5) , second sampling (Elongation – E2-E3, 2-3 nodes palpable =E2-E3),
The increase in herbage yield (kg DM ha-1) is related to the malate/sulphate-ratio in red clover and timothy leaves, respectively. The regression line that best describe the relation between the relative herbage yield and the malate/sulphate-ratio will be used as an indicator on possible yield decrease caused by low S-supply. The malat/sulphat-ratio is compared with other criteria for S-supply; concentration of sulphate and nitrate in timothy and red clover-leaves and the content of S, N/S- and P/S-ratio in red clover and timothy plants.
5.1.3. Adaptation of a Norwegian model for estimating K- release from soil to grass adjusted to organic farming
The model that is developed in the project ”Kalium som ressurs i dyrka jord” (Potassium resources in soil) {Øgaard, Krogstad, et al. 2002 3628 /id} will be adapted to organic farming by estimating the K release from the K-fractions K-AL and acid soluble K in the field trials. The K release from K-AL will be calculated from the difference between the K-AL values in spring and autumn. The change in K-AL during the growth season will be correlated with the K-AL value in spring minus the minimum level for K-AL to investigate if the change in K-AL, and thereby the K release from this fraction, might be predicted like shown in conventional leys (Øgaard et al. 2002). The K release from acid soluble K will be calculated as the amount of K in harvested plant material minus K release from K-AL and K applied in fertiliser. Calculated K release from acid soluble K will be correlated with the value of acid soluble K (K-HNO3 minus K-AL), to investigate if the release from acid soluble K might be predicted from acid soluble K.
2. Evaluation of different S-sources.
This will be planed in detail later.
I am very happy to get suggestions for S-sources relevant for organic farming.
So far we have the following suggestions. I am happy to get better suggestions
Seaweed, seaweedmeal, meat and bone meal (probably not so much S in this), urine, pig manure, vinasse.
Communication
Scientific articles are expected on the following subjects:
- The S-supply for organic leys in different places in Norway
- Testing the malate/sulphate ratio as a method to evaluate S-supply in organic ley
- Evaluation of different S-sources suitable for organic farming practise.
- Adaptation of a Norwegian model for estimating K- release from soil adjusted to organic farming
[SH1]This mighr go out
[SH2]Some authors have found that S is just as available in gypsum as in potassium sulphate and some have not. I suppose that depens on the Ca statusof the soil, the soil moisture and temeparure, perhaps the second suggestion for this hypothesis is a better one
[SH3]This will rewritten when we get the result from the soil analyses
[SH4]sjekk
[SH5]inspect if this small enough for chemical analyiss
[SH6]sjekk med lab hva som gjøres
[SH7]{regarding the method . note from Gustav Fystro – mars 2004. We started to calibrate NIRS for the prediction of clover content (grass/clover) some years ago. We early found that NIRS predictions were more reliable than visual estimates from experimental fields performed by the different persons responsible for the different trails. If we have known mixtures of clover and grass (artificially made from pure clover and pure grass) we found that our NIRS predictions off clover content (%) were within a standard error pf prediction (SEP) of about 3 %-units (using a 0 to 100% population). If we calibrate for a local population with a more narrow variation (e.g. 0-10% clover content) the SEP would be less than 1 %-unit (can not tell exactly). Some interfering problems might exist, as weeds (we do not know this very well). I think the problem with getting a representative sample is the most critical factor. }