Spatial methods and their applications to environmental and climate data

TL;DR

This paper reviews spatial and spatio-temporal statistical methods for environmental and climate data analysis, highlighting challenges like large datasets and non-stationarity, and discusses Bayesian hierarchical models with an application to Holocene land-climate interactions.

Contribution

It provides a comprehensive overview of spatial and spatio-temporal methods, including recent topics like large data handling and non-stationary covariance structures, with practical insights into Bayesian hierarchical modeling.

Findings

Bayesian hierarchical models offer flexible modeling options.

Handling large datasets remains a challenge in spatial statistics.

Application to LANDCLIM data demonstrates practical utility.

Abstract

Environmental and climate processes are often distributed over large space-time domains. Their complexity and the amount of available data make modelling and analysis a challenging task. Statistical modelling of environment and climate data can have several different motivations including interpretation or characterisation of the data. Results from statistical analysis are often used as a integral part of larger environmental studies. Spatial statistics is an active and modern statistical field, concerned with the quantitative analysis of spatial data; their dependencies and uncertainties. Spatio-temporal statistics extends spatial statistics through the addition of time to the, two or three, spatial dimensions. The focus of this introductory paper is to provide an overview of spatial methods and their application to environmental and climate data. This paper also gives an overview of several important topics including large data sets and non-stationary covariance structures. Further, it is discussed how Bayesian hierarchical models can provide a flexible way of constructing models. Hierarchical models may seem to be a good solution, but they have challenges of their own such as, parameter estimation. Finally, the application of spatio-temporal models to the LANDCLIM data (LAND cover - CLIMate interactions in NW Europe during the Holocene) will be discussed.