In the Northern Vietnamese Mountains, paddy fields are the main source of foods and an important indicator to define wealth in ethnic communities. This paper deals with an application of multilevel model to predict spatial patterns of paddy field productivity according to hierarchical levels such as individual fields, individual households and individual villages. The data on land use and rice productivity gathered in Sa Pa district (Lao Cai province) is used as a calibration and validation dataset. Totally, 60 households owning 136 paddy plots in four villages were interviewed. The results show that in Sa Pa, productivity of paddy fields results from human–environment interactions at different organizational levels of field, household and village. The variables collected at field and household levels play more important role in the effective use of paddy fields than the variables at village level, as they account for resp. 81.2%, 18.7% and 0.1% of the variance in productivity.

中文翻译：

---

多层模型在越南北部山区水田生产力影响因素分析中的应用

在越南北部山区，稻田是食物的主要来源，也是定义族裔社区财富的重要指标。本文应用多层次模型，根据各个领域，每个家庭和每个村庄的等级层次，预测稻田生产力的空间格局。萨帕地区（老街省）收集的土地利用和稻米生产率数据用作校准和验证数据集。总共采访了在四个村庄拥有136个稻田的60户家庭。结果表明，在萨帕州，稻田的生产力是由于田间，家庭和乡村的不同组织水平下人与环境的相互作用而产生的。在田间和家庭一级收集的变量在水田的有效利用方面比在乡村一级的变量起着更重要的作用，因为它们占了响应。生产率差异的81.2％，18.7％和0.1％。