Browsing by Author "White, Vivian George"

Now showing 1 - 1 of 1
  • Thumbnail Image
    Item
    Improving phosphate fertilizer recommendations using soil phosphorus buffer capacity and evaluation of various P extraction tests on a variety of South-African soils
    (Stellenbosch : Stellenbosch University, 2019-04) White, Vivian George; Hardie-Pieters, Ailsa G.; Raath, Pieter; Stellenbosch University. Faculty of Agrisciences. Dept. of Soil Science.
    ENGLISH ABSTRACT: Phosphorus (P) fertiliser recommendations are based on extractable P levels as measured by a variety of extraction tests and should incorporate factors such as depth, bulk density and the soils phosphorus buffering capacity. Internationally and in South-Africa a variety of extraction test are used with contrasting norms and crop norms are not available for all of these extraction tests. Therefore the first aim of this study was to examine the relative aggressiveness and relationships between the various P extraction tests using 49 soils with widely varying physicochemical properties. The relative efficiency as compared to Total P (acid digestion) of the extraction tests were as follows; 1:2 H2O < Olsen < Colwell < Bray I < Ambic 1 < Bray II < Mehlich III ≈ Citric acid. Bray II and Mehlich III were the only extractions test that correlated significantly (R2 = 0.78 & 0.76) with total P. Strong linear relationships were observed between Bray II, Olsen and Mehlich III extractions tests and based on the weighted standard error of measurement, direct conversions between these tests were most reliable. It is thus possible to convert with confidence from Olsen to Bray II and Bray I by using a factor of 5.20 and 3.88 respectively. For the conversion from Bray II to Mehlich III a factor of 1.10 can be used. Phosphorous buffering capacity (PBC) can be determined by a multiple-point sorption isotherm and through fitting the Freundlich equation or through a single-point isotherm method by making various assumptions. A subsample set of 10 soils varying in P sorbing capabilities was used, a large variance in parameter b was observed and ultimately two distinct groups with both a different value for b were determined. By equilibrating the samples with 1000 mg P kg-1 and by using the untransformed variation of the Freundlich equation and using the two distinct manually selected values for b (one for the low PBC soils and one for the high PBC soils) the best results were obtained. The single point estimate of PBC correlated significantly with PBC (Ozanne and Shaw, 1968) (Eq. 1 & 2). However, a large RMSE was observed and this predicted estimate of sorption is not reliable. These estimations utilise unrealistically high P levels and a 1:10 soil: solution, completely saturating the soil and allowing for most of the added P to react with the soil. This would be highly unlikely when incorporating P fertilizer into the soil thus a simple laboratory incubation method was investigated. A strong linear relationship existed between applied P and percentage extractable P by the three extraction tests used (Bray II, Mehlich III and Olsen) using the incubation method. It was found that at an application rate exceeding 150 mg P kg-1 the regression lines plateaued and that and the percentage extractable P at rates (100 – 150 mg kg-1) correlated highly significantly (R2 = 0.92 and 0.99 respectively) with the percentage extractable P derived from the slope of the regression line of applied P against extractable P. The influence of time also needed to be investigated and it was found that Bray II extractable P only significantly decreased after 1 month and that after 24 hours of equilibration the percentage extractable P remains fairly constant up to one month. Therefore it is possible to get an estimate of the amount of applied P that will be plant available with a single-application incubation method allowing for 24 – 72 h of equilibration. Lastly the Bray II and Olsen extractable P was correlated to plant response in high pH soils of the Northern cape planted to grapevine and citrus, no significant response was observed to the applied P fertilizer due to soil P already being at significant levels. Bray II was most sensitive in detecting applied P and excessive soil P levels, where Olsen was the least effective in detecting applied P. Therefore these soils do run the risk of reaching excessively high P levels which can in turn lead to P loss, micronutrient deficiencies and groundwater contamination.