"I have collected field data on heavy metal concentrations and, in 2024, developed multivariate regression models for each heavy metal using satellite imagery (such as Landsat or Sentinel) and vegetation indices, iron oxide, and clay minerals. I now intend to apply these models to the period from 2000 to 2023, for which I do not have field data.

My questions are as follows:

  • Is it scientifically valid to apply the model calibrated in 2024 to data from previous years (2000-2023) without having field data for those years?
  • What methods are suggested for evaluating the accuracy of such temporal extrapolation?
  • Are there reliable scientific sources (articles or reference books) that address this challenge and can be cited?"
  • More Javad Momeni Damaneh's questions See All
    Similar questions and discussions