面向农田建设土壤有机质改良成效评估的监测样点优化布设方法

Optimized layout method for monitoring sites for evaluating the effectiveness of soil organic matter improvement in farmland construction

关键词:  高标准农田建设  有机质改良成效评估  双重差分模型  空间样点布设

Key words:  high-standard farmland construction  soil organic matter improvement effectiveness assessment  difference-in-differences （DID） model  spatial sampling design

摘要:

【目的】 准确掌握农田土壤有机质改良成效信息对于支撑高标准农田建设工作具有重要意义。传统以制图或整体分布为目标的采样布局方法难以有效识别多源因素干扰影响下农田建设引起的土壤有机质改良成效。【方法】 文章提出一种面向农田建设土壤有机质改良成效评估的监测点优化布设方法。该方法结合因果推断思想通过模拟自然演化与高标准农田建设干预情景构建高标准农田建设场景，根据环境特征相似性提取建设区对应的非建设区样本对，并据此利用双重差分模型评估政策净效应实现监测点优化布设。【结果】 基于梨树县的应用结果显示，该方法所获得的土壤有机质改良成效估计均方根误差和平均绝对误差分别为0.007 20 g/kg和0.003 79 g/kg；样本量在50~300范围内逐步变化时评估精度保持稳定。【结论】 应用效果表明该监测样点优化布设方法能够显著提升土壤有机质改良成效监测准确性和稳健性水平，可为区域高标准农田建设土壤质量改良信息高效获取提供科学可靠的技术支撑和实现途径。

Abstract:

[Purpose] Accurately quantifying the improvement in farmland soil organic matter （SOM） is of critical importance for supporting high-standard farmland construction initiatives. Traditional sampling layout approaches primarily designed for mapping or overall spatial distribution analysis are often inadequate for effectively identifying SOM improvement driven by farmland construction under the influence of multiple environmental factors.[Method] To address this limitation，this study proposed an optimized monitoring point allocation method specifically designed for evaluating SOM improvement associated with farmland construction. Grounded in causal inference principles，the proposed method constructed high-standard farmland scenarios by simulating natural evolution and construction intervention processes. By incorporating environmental feature similarity matching，non-construction samples corresponding to construction areas were identified to form comparable pairs. A difference-in-differences （DID） model was then employed to estimate the net construction effect，thereby achieving optimized monitoring point allocation.[Result] Application in Lishu County，Jilin Province demonstrated that the proposed method yielded a root mean square error （RMSE） of 0.007 2 g/kg and a mean absolute error （MAE） of 0.003 79 g/kg in estimating SOM improvement. Furthermore，when the sample size varies from 50 to 300，the evaluation accuracy remained stable.[Conclusion] These results indicate that the optimized monitoring point allocation method significantly improves the accuracy and robustness of SOM improvement assessment，providing a scientifically reliable and efficient technical pathway for acquiring soil quality improvement information in regional high-standard farmland construction.