Scaling up or scaling down: the use of foliage and soil information for optimising the phosphate nutrition of radiata pine

Abstract

Fertilising with phosphate is important in New Zealand radiata pine plantations. Phosphate fertiliser trials have provided functions for determining potential growth gains from stands based on foliar P status. We used the national foliage database to determine the proportion of forest stands likely to be deficient in NZ. This gave a statistical representation but without a spatial reference. We used a nested design to test the spatial representation of foliar nutrient status and responsiveness at varied map scales. First we used the national soil map (1:10 6 scale) to show where P deficiency, and hence response, was most likely to occur. It was apparent that this representation was not acceptable for any operational scale planning. General trends in foliar status could be seen but there was no way of defining where foliage sampling should be concentrated within soil units, leading to inefficient sampling strategies. We then concentrated on the pumice, allophanic and podsol soil orders in the central North Island and used foliage data associated with 1:100,000 scale soil series map units. The spatial variation in likelihood of P deficiency and responsiveness was much clearer at this scale and more efficient foliar sampling programmes could be designed. At the most intensive scale (1:10,000) there were further gains to be made in foliage sampling efficiency and identification of responsive sites in some, but not all cases, and this depended on the soil pattern. We then tested whether spatial analysis of foliage data without reference to soil was better than using soil series as a defining class. Semivariograms constructed for foliar P concentration indicated that errors involved with predicting foliar P concentration were no less than if soil series was used as a predictor of foliar P status and responsiveness. We concluded that 1:100,000 soil series maps were adequate for general nutritional management planning and definition of areas likely to be responsive. However, for really precise work 1:10,000 maps would give additional definition for P fertilisation operations and allow us to refine fertiliser prescriptions at the within stand scale in some cases.