Simulating cropping sequences using earth observation data
Simulating cropping sequences using earth observation data
复制标题
使用地球观测数据模拟种植序列
DOI：
10.1016/j.compag.2021.106330
复制
发表时间：
2021
影响因子：
8.3
通讯作者：
Sharp R
中科院分区：
文献类型：
--
作者：
Sharp R
基于模型的农业系统研究往往依赖于分析师定义现实的作物序列。这通常涉及到依赖于一些在基线场景中使用的“典型旋转”。然而，这些可能无法解释一个地区耕作方式的差异，因为农民种植哪种作物的决策受到经济、环境和社会驱动因素的综合影响。我们描述和测试一种方法，用于生成随机实现的合理作物序列的基础上观测到的数据量化的地球观测。我们的方法将作物分类数据与一系列作物管理规则相结合，这些规则反映了农民遵循的建议（例如减少作物病虫害的机会）。我们采用这种方法来生成特定于地区和土壤类型的作物序列。这证明了该方法如何适用于生成感兴趣的研究区域的典型作物序列。
Model-based studies of agricultural systems often rely on the analyst defining realistic crop sequences. This usually involves relying on a few ‘typical rotations’ that are used in baseline scenarios. These may not account for the variation in farming practices across a region, however, as farmer decision making about which crops to grow is influenced by a combination of economic, environmental and social drivers. We describe and test an approach for generating random realisations of plausible crop sequences based on observed data as quantified by earth observation. Our approach combines crop classification data with a series of crop management rules that reflect the advice followed by farmers (e.g. to reduce the chance of crop-pests and disease). We adapt the approach to generate crop sequences specific to regions and soil type. This demonstrates how the method can be adapted to generate crop sequences typical of a study area of interest.