期刊
CROP & PASTURE SCIENCE
卷 73, 期 6, 页码 636-651出版社
CSIRO PUBLISHING
DOI: 10.1071/CP21627
关键词
decision trees; digital soil mapping; pedotransfer function; precision agriculture; Random Forest; soil constraints; soil water; yield potential
资金
- University of Sydney
In this study, machine learning and digital soil mapping techniques were used to map the depth of constraints in a land area of approximately 80,000 hectares in northern New South Wales, Australia. The study found that constraints were present in the upper 1.2 meters of the soil profile in 54% of the study area, resulting in an average reduction of 50 mm in available water capacity (AWC) and an average yield loss of 1.1 t/ha for wheat and 0.8 bales/ha for cotton. Sodicity was identified as the most limiting constraint to yield in the study area.
Context The depth-to a constraint determines how much of the soil profile, and the water it contains, can be accessed by plant roots. Information describing the impacts of soil constraints on available water capacity (AWC) and yield is important for farm management, but is rarely considered in a spatial context. Aims and methods The depth-to three yield-limiting constraints (sodicity, salinity, and alkalinity) was mapped across similar to 80 000 ha in northern New South Wales, Australia using machine learning and digital soil mapping techniques. Soil AWC was calculated using soil data and pedotransfer functions, and water use efficiency equations were used to determine potential yield loss due to the presence of soil constraints. From this, the most-limiting constraint to yield was mapped. Key results One or more constraints were found to be present across 54% of the study area in the upper 1.2 m of the soil profile, overall reducing the AWC by similar to 50 mm and potential yield by an average of 1.1 t/ha for wheat and 0.8 bales/ha for cotton. Sodicity (Exchangeable Sodium Percentage > 15%) was identified as the most-limiting constraint to yield across the study area. Implications The simplification of multiple sources of information into a single decision-making tool could prove valuable to growers and farm managers in managing soil constraints and understanding important interactions with available water and yield.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据