科学普及
当前位置: 首页 >> 科学研究 >> 科学普及 >> 正文
    知识科普|基于随机森林方法研究长期保护性耕作对欧洲中心土壤侵蚀的影响
    2021年12月14日 18:00

水土流失是威胁农业可持续发展的重要因素之一。作者通过在匈牙利西南部进行16年田间定位试验,探究保护性耕作对田间径流(RO)和土壤流失(SL)的影响,并采用随机森林建立模型确定影响田间径流与土壤流失的主要因素。

Soil erosion is one of the most important factors threatening sustainable agricultural development. The authors investigated the effects of conservation tillage on runoff (RO) and soil loss (SL) through a 16-year field study in southwest Hungary and used a random forest-based model to identify the main factors affecting RO and SL.

研究发现,与翻耕相比(PT),保护性耕作方式(CT)下土壤水稳性团聚体含量提高10.4%,有机质含量增加0.5%,提高了蚯蚓丰度;同时降低了75%的地表径流和95%的土壤流失。通过将影响田间径流和土壤流失的相关因素(耕作方式、作物种类、植被覆盖、发生时间、降水量、降水强度、降水持续时间)采用随机森林模型进行建模,结果显示:耕作方式是影响田间地表径流和土壤流失的最重要因素

It was found that in the soil, the water-stable aggregates content increased by 10.4%, organic matter content increased by 0.5%, and earthworm abundance increased under conservation tillage (CT) compared to ploughing tillage (PT); while 75% of RO and 95% of SL were reduced. By modeling the relevant factors affecting RO and SL (tillage practices, crop types, vegetation cover, time of occurrence, precipitation amount, precipitation intensity, and precipitation duration) and using a random forest-based model, the results showed that the tillage practices were the most important factors affecting field RO and SL.


与翻耕相比,保护性耕作可以减少地表径流和土壤流失。虽然两种耕作方式年均土壤流失量均小于2 t/ha,但翻耕方式下有5个年份土壤流失量超过安全阈值。

CT reduces RO and SL compared to PT. Although the average annual soil loss was less than 2 t/ha for both tillage practices, SL exceeded the safety threshold in five years under PT.

4种建模中,影响田间地表径流和土壤流失的最重要因素均为耕作方式。对于所有作物类型,降水强度是影响土壤流失的第二重要因素;降水强度、降水持续时间以及作物类型的联合作用是影响地表径流的重要因素。

The most important factor affecting RO and SL in the field was tillage practices in all four modeling types. For all crop types, precipitation intensity was the second most important factor; the combined effect of precipitation intensity, precipitation duration, and crop type was also an important factor.

对于玉米来说,降水持续时间、植物覆盖度以及降水强度是解释地表径流的重要因素;而对于土壤流失,植物覆盖度、发生时间以及降水强度起决定性作用。

For corn, precipitation duration, plant cover, and precipitation intensity were important factors in explaining RO; while for SL, plant cover, time of occurrence, and precipitation intensity played a decisive role.

因此,保护性耕作可以有效缓解水土流失。随机森林模型可以运用于预测水土流失。

Therefore, CT can effectively mitigate soil erosion. Random forest-based model can be applied to predict soil erosion.




投稿者:高峰 博士研究生

审核导师:王宁 教授 |王慧 助理研究员

 

南海海洋资源利用国家重点实验室(海南大学)
  • 地址:海南省海口市美兰区人民大道58号海南大学 (邮编:570228)
  • 电话:0898-66292367 传真:0898-66292367
  • 邮箱:mru@hainanu.edu.cn mruzp@hainanu.edu.cn(招聘专用)
  • 版权所有 @2017 未经授权禁止转载、摘编、复制或建立镜像.
  • 美工支持/中旗网络  后台管理
  • 公众号