Applying Geospatial Semantic Array Programming for a Reproducible Set of Bioclimatic Indices in Europe

TL;DR

This paper presents a reproducible geospatial methodology using semantic array programming to derive 12 bioclimatic indices for Europe, facilitating ecological zoning and ecological niche modeling with open data and tools.

Contribution

It introduces a reproducible workflow combining GIS and semantic array programming to generate bioclimatic indices from global open datasets for Europe.

Findings

Derived 12 bioclimatic covariates for Europe using open data.

Workflow ensures reproducibility and expandability of bioclimatic indices.

Indices are made available as open data for further ecological studies.

Abstract

Bioclimate-driven regression analysis is a widely used approach for modelling ecological niches and zonation. Although the bioclimatic complexity of the European continent is high, a particular combination of 12 climatic and topographic covariates was recently found able to reliably reproduce the ecological zoning of the Food and Agriculture Organization of the United Nations (FAO) for forest resources assessment at pan-European scale, generating the first fuzzy similarity map of FAO ecozones in Europe. The reproducible procedure followed to derive this collection of bioclimatic indices is now presented. It required an integration of data-transformation modules (D-TM) using geospatial tools such as Geographic Information System (GIS) software, and array-based mathematical implementation such as semantic array programming (SemAP). Base variables, intermediate and final covariates are described and semantically defined by providing the workflow of D-TMs and the mathematical formulation following the SemAP notation. Source layers to derive base variables were extracted by exclusively relying on global-scale public open geodata in order for the same set of bioclimatic covariates to be reproducible in any region worldwide. In particular, two freely available datasets were exploited for temperature and precipitation (WorldClim) and elevation (Global Multi-resolution Terrain Elevation Data). The working extent covers the European continent to the Urals with a resolution of 30 arc-second. The proposed set of bioclimatic covariates will be made available as open data in the European Forest Data Centre (EFDAC). The forthcoming complete set of D-TM codelets will enable the 12 covariates to be easily reproduced and expanded through free software.