Preface

Who this book is for

This repository contains a review of Global in-situ Observations & Measurements networks and data sets, and a number of compiled and analysis ready point (spatial) data sets that can be used for training global spatial prediction models or similar analysis. It is has been developed for the purpose of building the www.OpenLandMap.org data sets, which is one of the main outputs of the Open-Earth-Monitor Cyberinfrastructure project.

The main objective of this compendium is to support a more reproducible and a more collaborative global modeling and analysis of environmental and Earth System Science variables. We believe that the key to making environmental data more trustworthy is by increasing its FAIRness (findability, accessibility, interoperability and reusability), and by enabling open development communities that freely exchange code and help make better open source software.

Compiled data sets

We provide code and examples of how to generate so-called Analysis-Ready training data sets. Some minimum conditions for a data set to be analysis ready include:

• It requires no special pre-processing to remove artifacts, harmonize values within columns, bind or subset data;
• It comes with extensive metadata so that there is no mistake about how was the data collected, prepared and distributed and by whom;
• It is ready in some standard format e.g. GeoPackage; for large data sets we recommend using (open) cloud-native data formats to distribute data either Geoparquet, Flatgeobuf or similar;