Potassium releasing and supplying power of selected yellow grey earth soils of New Zealand : a thesis presented in partial fulfilment of the requirements for the degree of Doctor of Philosophy in Soil Science at Massey University

The supply of soil potassium (K) to New Zealand pastures is currently being assessed by the quick test K (QTK) and reserve K (I(.,) methods, which measure soil exchangeable K (Kex) and non-exchangeable K (K..ex)' respectively. QTK is based on a routine soil test and I(., is an assigned estimate appropriate to the soil group. No consideration is given to the variations of the K..ex supply within a soil group. The objective of this research was to examine the K releasing and K supplying power of selected soils from the yellow-grey earth (YGE) group. A wide variation was observed in the measured I(., values of the YGE soils in the North and South Islands. A glasshouse experiment showed that the supply of K..ex to ryegrass grown on the 13 North Island YGE soils ranged from 0-41 mg 100 g.! and that of the 6 South Island YGE soils ranged from 3-35 mg 100 g.!. The experiment also showed that there were lower levels of K..ex supply in the pasture sites, compared to the virgin sites with respect to the South Island YGE soils. These results have implications to the use of the soil group concept which is used to estimate K..ex supply in the Computerised Fertilizer Advisory Service (CFAS) K model, currently used by AgResearch. In a laboratory study, the threshold K levels in terms of K concentration and the activity ratio in the equilibrated soil solution, Kex' and the amount of specifically held K were determined, in order to explain the variations in Kuex supply. The threshold K levels were not related to the Knex release and supply. The uptake of K by ryegrass was at best poorly to moderately correlated with the K extracted by current methods of determining K releasing power viz, QTK and 1(.,. The highest simple correlation was obtained from an improved acid-extractable K procedure (r = 0.96; P < 0.01). The differences in the Kuex uptake by ryegrass from various soils were better explained by a simple method of determining soil Kuex i.e., step K, than by the existing I(., method. A multiple regression equation with QTK and step K as independent variables explained 96 % of the variation in total K uptake among soils. iii On the basis of K..e x taken up by ryegrass in the glasshouse experiment, the 19 soils in this study were broadly grouped into two categories (i) soils with step K values of less than 35 mg 100 g'! and a K" range of 8-10 mg 100 g'! and (ii) soils with step K values greater than 35 mg 100 g'! and a K" range of 12-19 mg 100 g'!. Selected soils were fractioned into sand, silt, and clay separates and acid-extractable K levels of the fractions were measured. There was a wide range in the acid­ extractable K levels among the soils for the same size fraction e.g., clay, and for different size fractions within the same soiL When weighted according to the particle size distribution of the soil, the sand was found to contribute 4-45 %, silt 1040 %, and clay 15-85 % of the K released by the sum of the 3 separates, using the improved acid extraction method. In all the soils, the clay separate released the most K per unit weight. An agar pot trial technique was developed to measure the K supplying power of the soil separates. Although on a unit weight basis the clay separates showed a much greater activity than the other separates on a weighted basis, the contributions of sand and silt separates to the total K uptake of Marton (38 %), Matapiro (41 %), and Wharekaka (25 %) soils was of considerable importance. The results demonstrated that the role of sand and silt separates deserve more consideration in estimating potential K releasing and supplying power than has hitherto bee n the case. The study also attempted to relate K..ex release and supply to the soil mineralogy. Although the gross mineralogy of the 19 soils was similar, differences in the K..ex release and supply could be related to subtle differences and gradual changes in the clay mineralogy. The XRD patterns of the clays with a K" range of 8-10 mg 100 g'! of soil differed from those with a K" range of 12-19 mg 100 g'! of soiL The latter group of clays contain more K bearing minerals than the former group. The practical implications of the measured differences in � values (�x supply) within the YGE soil group were dealt with. The measured �x supply in the North Island YGE soils ranged from 20-40 kg ha·! yr·!, whereas the expected �x supply based on an assigned K.: value is 30 kg ha·! yr"!. The difference between the expected and the measured K..ex may be sufficiently economically significant as to invalidate applying a single K.: value to a soil group. Possible improvements to the soil K supply component of the CFAS K model were suggested, particularly that step K values should replace K.: in the K supply model.