Statistical modelling in time and space for environmental monitoring

The SLU Centre for Statistics (Statistics@SLU) invites you to a Two-Day Workshop on statistical analysis of environmental monitoring data

Start date: 4 March 2027

End date: 5 March 2027

Venue: Zoom-link will be available at this page later

Language: English

Organiser: Statistics@SLU

Location: Online

Basic statistical models often assume independent observations. However, when monitoring the environmental observations are intrinsically correlated in time, in space of both. The workshop gives an overview over methods and concepts used in environmental monitoring and discusses pros, cons and limitations.

Programme

Time:

4 March 2027  9:15 - 12:00

5 March 2027 9:15 – 12:00

Content:

  • Temporal and spatial autocorrelations and their estimation
  • Analysis of time series:
    • Data-driven: Generalized additive models with autocorrelated errors (GAMM)
    • Non-parametric: Mann-Kendall, Theil-Sen slope
    • Parametric: Generalized linear models with autocorrelated errors (GLMM)
  • Analysis of spatial data:
    • Parametric: GLMM with spatially correlated errors
    • Data-driven: GAMM with spatially correlated errors, geographically weighted regression models (GWR)
    • Machine learning: spatial random forests (RF-GLS)
  • Other concepts connected to analysis of environmental monitoring data
    • Seasonality, outliers, gaps in data,…
    • Analysis of values below a reporting limit
    • Spatially or temporally varying coefficient models

Leader:

Claudia von Brömssen,  Department of Energy and Technology, Unit of Applied Statistics and Mathematics, SLU 

A Ph.D. course might potentially be tied to this workshop. Contact claudia.von.bromssen@slu.se for more information