Forum
Measurement Uncertainty in Ecological and Environmental Models

https://doi.org/10.1016/j.tree.2020.07.003Get rights and content

In many applied cases of ecological and environmental modeling there is sizeable variation in the independent variables as a result of measurement and sampling errors. This uncertainty may lead to biased predictions. It is possible to avoid this problem by increased sampling and by modeling the errors using hierarchical modeling.

Section snippets

Stochastic Modeling and the Focus on Realism

Predictions based on ecological and environmental models are often used to advise administrative and political systems, and thus play an important role in making political decisions on environmental issues [1]. Most directly, this is seen in fish stock models, which have an immediate influence on the politically determined fish catch rates [2].

To strive for precision and realism, many ecological and environmental models have been developed that are functionally complex and have many input

Biased Predictions

As demonstrated in Box 1, measurement and sampling errors may lead to biased predictions, and it is clear from Figure I that the predictive performance of the model is weakened at the borders of the domain of the independent variable. In the case presented, the predictions were heavily biased downwards (see Table I in Box 1).

If each of the independent observations is subsampled ten times, in other words xi is now determined by a sample of ten independent observations, then the data can be

Recommendations

The main conclusion of these simulations is that it is important to reduce any sizable measurement and sampling errors in the independent variables by independent subsampling of the variables. In principle, it is then possible to use traditional regression techniques to fit the mean values to the model. However, if the fitted models have several independent variables that all have variable measurement and sampling errors and/or a variable degree of subsampling, then some independent variables

References (11)

  • C. Damgaard

    Modelling pin-point cover data of complementary vegetation classes

    Ecol. Inform.

    (2015)
  • National Research Council

    Models in Environmental Regulatory Decision Making

    (2007)
  • W.H. Aeberhard

    Review of state-space models for fisheries science

    Annu. Rev. Stat. Appl.

    (2018)
  • M.R. Evans

    Modelling ecological systems in a changing world

    Philos. Trans. R. Soc. B

    (2012)
  • R.J. Carroll

    Measurement Error in Nonlinear Models: A Modern Perspective

    (2006)
There are more references available in the full text version of this article.

Cited by (0)

View full text
© 2020 Elsevier Ltd. All rights reserved.