资源环境科技发展态势分析平台

Soil fertility is one of a number of environmental and biotic drivers of Pinus radiata (D. Don) productivity in New Zealand's planted forest estate. However, its importance as a driver of productivity at a national scale is uncertain. This paper uses New Zealand's planted forest inventory plot network, which encompasses a wide range of planted forest environments in New Zealand, to quantify climatic, soil, and biotic drivers of P. radiata productivity. The P. radiata index of site productivity used in this study, the 300 Index, is defined as the mean annual increment in stem volume under bark at age 30 years, of stands growing at a final crop stocking of 300 trees per hectare, and thinned and pruned following a standardised regime. This site productivity index was found to be significantly related to climate (30 year normalised average temperature and rainfall), solar radiation (corrected for slope, horizon and aspect using a digital terrain model), tree health, and soil fertility variables. After adjusting for the effects of solar radiation, temperature and rainfall, site productivity across the New Zealand planted estate was found to be negatively related to the soil C/N ratio and positively related to the size of the soil organic matter pool. The size of the organic matter pool could be represented in the regression model by either the total carbon, total nitrogen, or organic phosphorus concentration of the topsoil. Of these three soil variables, soil total nitrogen was statistically the most significant variable in the model, when analysed in conjunction with the soil C/N ratio. Afforested pastoral farmland previously treated with phosphatic fertilisers to stimulate nitrogen-fixation by leguminous species had soils with a lower C/N ratio and a higher organic matter content. This contributed to an 18% gain in P. radiata productivity compared to land without an agricultural land use history. Having established a relationship between the P. radiata productivity index and climate and soil fertility key drivers, we also consider methods for improving regression model accuracy.