Scholarlink

<p>To test the concept of automated pest surveillance for smart agriculture, infrared (IR) sensors were deployed on a rectangular grid in a commercial California citrus orchard to monitor Argentine ant activity. Since ants use irrigation pipes as highways, the IR sensors were placed along the irrigation pipes, taking 1-min ant counts every hour for 10 days, yielding spatially correlated ant count time series. A Bayesian spatiotemporal model framework that accounted for weather covariates, diurnal cycles, and pipeline and sensor random effects was developed to understand the sources of variation and the spatiotemporal correlation in the recorded ant count data. The proposed Bayesian model was fitted using Markov chain Monte Carlo via the NIMBLE package in R. The model’s numerical performance was tested through simulation studies, and real data analysis results shed new light on understanding ant activities in orchard environments. The proposed methods can be widely used to model spatially correlated time series data collected by sensors, which are increasingly used in agricultural, biological, and environmental applications. Supplementary materials accompanying this paper appear online.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Development of Bayesian Spatiotemporal Models to Analyze Ant Count Data Collected by Automated Infrared Sensors in Commercial California Citrus Orchards

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年7月25日

查看全文
  • Yinghuai Yi,
  • Ivan Milosavljević,
  • Mark. S. Hoddle,
  • Yehua Li

摘要

To test the concept of automated pest surveillance for smart agriculture, infrared (IR) sensors were deployed on a rectangular grid in a commercial California citrus orchard to monitor Argentine ant activity. Since ants use irrigation pipes as highways, the IR sensors were placed along the irrigation pipes, taking 1-min ant counts every hour for 10 days, yielding spatially correlated ant count time series. A Bayesian spatiotemporal model framework that accounted for weather covariates, diurnal cycles, and pipeline and sensor random effects was developed to understand the sources of variation and the spatiotemporal correlation in the recorded ant count data. The proposed Bayesian model was fitted using Markov chain Monte Carlo via the NIMBLE package in R. The model’s numerical performance was tested through simulation studies, and real data analysis results shed new light on understanding ant activities in orchard environments. The proposed methods can be widely used to model spatially correlated time series data collected by sensors, which are increasingly used in agricultural, biological, and environmental applications. Supplementary materials accompanying this paper appear online.