The first round of projects under the UKCEH-BioSS Methodological Development Framework has been completed. This framework is a collaboration agreement set up in 2022 between UKCEH and BioSS, seeking to develop the availability and accessibility of quantitative methods through small methodological projects, for exploiting data collected and available across the breadth of UKCEH science areas. 

Understanding the status of natural or social systems and how they change over time, relies heavily on the scale at which data are available (the “support”) and whether this coincides with the scales at which the underlying natural or social processes operate. For example, measuring variables like biodiversity or crop health at the level of county areas may hide important variations between local areas, such as farms. Conversely, understanding these variables at fine scales (say, at the level of a farm), may require considering the broader regional and historical context, such as the environment it interacts with, or past management over short and longer timescales. Ignoring the multiplicity of scales can lead to researchers ignoring useful information in the data and may lead to incorrect conclusions. Integrating information across scales has been identified as a very general problem for researchers seeking to analyze spatial and temporal data. 

However, statistical methods for dealing with these issues tend to require specialist software that often lacks the user-friendliness and flexibility required by the empirical research community.

The first round of the UKCEH-BioSS Methodological Development Framework was thus dedicated to the general issue of handling changes in both biological and sampling processes across spatial or temporal scales. In particular, two specific questions emerged as priorities for either developing new methodology or to make existing advanced models more accessible: 1) models for handling changes of support in space and time, i.e. when system variables are observed at different locations or scales, in space or in time across data sources; 2) general methods for identifying the extent of the "zone of influence" of environmental and ecological predictors on processes of interest, in space, time, or space & time simultaneously. These gave rise to two distinct projects: 

• Project 1: “Data Fusion With Change Of Support”
• Project 2: “Estimating Effect Size Variation Across A Zone Of Influence” 

The modelling techniques for both projects were implemented with Generalized Additive Models (GAMs), using the popular ‘mgcv’ package in R.

The outputs of the two projects were delivered in a 2-day online workshop, in November 2023. Slides and tutorial vignettes are available on the workshop website. In support of the delivery workshop, we have put together this bespoke “Introduction to Generalized Additive Models” online course (2 days) which covers the key pre-requisites.

The code for the 2023 projects of the UKCEH-BioSS Methodological Development Framework is available here: https://gitlab.bioss.ac.uk/ukceh-bioss-framework/outputs.

BioSS contact: Thomas Cornulier thomas.cornulier@bioss.ac.uk. UKCEH contact: Pete Henrys pehn@ceh.ac.uk