【摘要】 土壤养分累积是多个尺度自然和人为因素共同作用的结果,在空间上呈现出多尺度分异性。定量解析这种多尺度分异规律及其驱动力对于养分管理至关重要。基于珠江三角洲地区417个表层土壤样品和6大类环境变量(气候、成土母质、土壤理化性质、地形/植被覆盖、距污染源的距离、社会经济),采用多元因子克里格方法定量解析土壤总氮、总磷、总钾、碱解氮、有效磷和有效钾的多尺度空间变异及其环境驱动力。结果显示,研究区土壤养分相对丰富,尤其耕地和园地养分含量较高,过量施肥以及养分利用率低是主要原因。研究区土壤有效磷为高度变异(变异系数为108%),其他养分质量分数均为中等变异(变异系数介于44%—78%),且养分高值区主要位于研究区中部(广州、佛山、珠海、惠州中西部和东莞西部)。因子克里格将土壤养分的空间变异分解为块金尺度、短变程尺度(26 km)和长变程尺度(90 km)3个独立尺度空间变异之和。土壤养分的空间变异格局及环境驱动力因尺度而异。短变程尺度,土壤养分的空间变异性更强,高值分布区较为分散,土壤养分质量分数主要受控于海拔和人为活动(如工业排放、社会经济活动)。长变程尺度,土壤养分的高值分布区较为聚集,土壤养分质量分数更多受控于自然因素(气候和成土母质)。基于"特征-机制-管理"思路,该研究提出土壤养分多尺度管理概念模型,根据不同尺度肥力区划及其驱动力制定不同的养分管理策略。更多还原

【Abstract】 The accumulation of soil nutrients is the outcome of the combined effects of natural and anthropogenic factors at multiple scales. Quantifying the spatial variability of soil nutrients and its driving forces at different scales is critical for nutrient management. This study applied a multivariate Factorial Kriging method to explore the spatial variability of soil nutrients(i.e. total nitrogen, total phosphorus, total potassium, alkali hydrolysable nitrogen, available phosphorus and available potassium) and its environmental drivers at multiple scales. We collected 417 surface soil samples and used six types of environmental factors(i.e., climate, soil physiochemical properties, the normalized difference vegetation index, the distance to pollutant sources, and socioeconomic factors) for attribute analysis. We found that soil nutrient contents in the study area were relatively high, especially for the farmland and orchard mostly due to excessive fertilization and low nutrient utilization efficiency. The available phosphorus was highly variable across space as indicated by the coefficient of variation of 108%, while other soil nutrients were moderately variable with the coefficient of variation between 44% and 78%. The areas with high nutrient contents were mainly located in the middle of the study area(i.e., Guangzhou, Foshan, Zhuhai, central and western Huizhou, and western Dongguan). Factorial Kriging decomposed the total variance of soil nutrients into a sum of nugget effect and variations of short-range(26 km) and long-range(90 km) scales. The spatial variability of soil nutrient distribution and environmental drivers varied with the scales examined. The heterogeneity of soil nutrients decreased with from a small scale to a large one. At the smaller scale(26 km), the spatial variability of soil nutrients was higher and the areas with high nutrient values were more scattered. The spatial variations in soil nutrient contents at this scale(26 km) was probably dominated by altitude and human activities such as industrial emissions and socioeconomic activities. At the long-range scale(90 km), areas with high nutrient contents were more clustered. The spatial variations of soil nutrient contents at this scale was dominated by natural factors(i.e., climate and parent materials). Based on the principle of "soil characteristics-processes-management", we developed a multi-scale conceptual model for soil nutrient management in the study area. We should formulate different nutrient management strategies at different spatial scales according to the zoning of soil nutrient contents(high vs. low) and their driving forces.更多还原