Changes in the phosphorus status of soils and the influence on maize yield
Schmidt, Christiaan Jan Jacob
MetadataShow full item record
The principle objective of this study was to obtain P fertilizer guidelines for large maize producing areas west of the Drakensberg mountains according to a similar approach previously used in KwaZulu-Natal based on the sufficiency concept of soil extractable P. Data from 14 different P fertilizer trials at various localities in the Free State, Gauteng, Mpumalanga and North West provinces was used. Different phosphorus treatments were applied for all trials in order to establish differences in extractable soil P levels which were expected to have corresponding effects on maize yield. Long-term rainfall varied from 990 mm per annum for the Athole trial in the eastern maize producing region to 494 mm per annum for the Wolmaransstad trial in the western region. The duration of trials varied between one and nine seasons. Firstly, simple regression equations with high R²-values were obtained for relationships between Ambic 1 and Bray 1 extractions over soils, but since it was demonstrated that relationships for different soils differed significantly from each other the use of these equations may result in a very high degree of inaccuracy with respect to P fertilizer recommendations. Soil properties had a significant effect on the efficacy of the two extractants. Furthermore, slopes of relationships between Ambic 1 and Bray 1 could be predicted by using exchangeable Ca in simple regression relationships (R²-values of between 80 and 83%). Secondly, it was established that P requirement factors (PRF's) cannot be obtained over soils (R²-values varied between 10 and 54 %), but rather for different soils separately (R²-values varied between 75 and 99%). Differences between the PRF's in total soil volumes (1.7 to 63.2 for Ambic 1 concentrations and 0.8 to 27.3 for Bray 1 quantities) indicated that the soils used in this study differed in their behavior to applied P. Phosphorus requirement factors could be predicted by a simple regression equation using degree of leaching based on the clay content as input parameter (R²-values between 60 and 78 %) as well as six multiple regression equations using either one of exchangeable Ca, Mg, K, silt content or degree of leaching based on the clay content (R²-values between 52 and 99 %) as input parameters. The implementation of any of these regression equations should be practical since all the parameters are usually included in standard analysis. However, the simple regression with degree of leaching based on the clay content appears to be an obvious option above the multiple regression equations since it is based on five input variables, i.e. exchangeable Ca, K, Mg, Na and clay content. Lastly, threshold extractable P values were derived for 10 out of the 14 localities that have been included in this study with varying R²-values. These threshold extractable P values were related to soil properties and it was found that the degree of leaching and silt-plus-clay content were the parameters that explained most of the variation. However, it was decided to explore only the relationships between threshold extractable P values and silt-plus-clay contents in more detail. By excluding data from two localities of which the topsoil contained lime, the R²-values of the mentioned relationships improved substantially so that threshold extractable P values could be derived from the silt-plus-clay content range of the other eight localities. For example the threshold extractable soil P concentrations based on Bray I for the total soil volume to obtain 90 % relative yield varied from 33.5 mg kg-¹ at 13 % silt-plus-clay to 14.6 mg kg-¹ at 60 % silt-plus-clay. These P (Bray 1) thresholds are much higher on the sandy soils than the value of 19 mg P kg:' (Bray 1) for 95 % relative yield currently in use according to existing guidelines. This may not necessarily imply that more P fertilizers will be sold according to higher soil P thresholds obtained in this study, since the corresponding soil sampling procedure also measures more residual P from enriched zones over rows where P fertilizer was band placed. The soil sampling procedure according to existing guidelines excludes sampling from these zones.