Chapter 1: Introduction#

Researchers and practitioners are interested in how the systems they work on will be affected by climate change. For many applications, users require highly detailed climate information about the future, but the climate projections produced by raw climate model output usually do not meet these needs. The UTCDW Guidebook will introduce you to important concepts in climate science required to understand how climate projections are made, and teach you practical skills regarding how to make climate model output useful for assessing the impacts of climate change.

Guidebook Material#

Before working through the material in this guide, we encourage users new to climate science and analysis with climate data to work through some of the course modules from an online series developed as a predecessor to the UTCDW Guidebook: Engineering in a Changing Climate: A Transdisciplinary Workshop Series for Engineering and Climate Science Students. Module 1 of the aforementioned resource (Introduction to Climate Science and Modelling) will introduce basic concepts in climate science and climate modeling. These topics will be expanded upon in Chapter 2 of this Guidebook. Chapter 3 of the Guidebook will demonstrate to the user how to acquire and explore climate data, both from observations and climate model simulations, using the Python programming language.

Module 2 of the Engineering in a Change Climate series (Regional Downscaling) discusses the basics of post-processing climate change projections to make them appropriate and useful for local-scale applications (i.e. for studying a particular region, city, or location). This is called downscaling (the “D” in UTCDW), and it is crucial for any climate change impact study. Chapter 4 of the Guidebook will expand upon the e-Learning module by discussing topics including methods for statistical downscaling in practice, and which methods are suitable for different applications.

The latter portion of the UTCDW Guidebook will introduce the “W” in UTCDW. Guidebook Chapter 5 will describe the decisions that must be made during study design, and help the user develop a workflow for acquiring, processing, and analyzing climate change projection data tailored for their study application. Finally, Chapter 6 will lay out examples of implementing this workflow for several different downscaling methods.

A Note About Computational Resource Requirements#

The volume of data to be processed as a part of a climate impact study can be very large. Multiple decades of global climate model data for a single variable at daily time frequency (e.g., daily mean air temperature at 2 m height) can total tens or even hundreds of gigabytes. Some data hosting services do not support spatial sub-setting before downloading the data, so for many use cases you will need significant data storage resources, if not significant computing (processor and memory) resources as well. This guide will assume you have access to such resources, either via your research group’s in-house cluster, an HPC cluster such as SciNet at the University of Toronto, or a cloud computing service such as Amazon Web Services. Certain free-to-use climate analytics platforms are available, such as PAVICS, but this guide aims to be independent of the computing platform you intend to use and does not provide support or guidance regarding their use.

A Note about Statistics#

The downscaling and analysis methods described and implemented in the UTCDW Guidebook make use of fundamental concepts in probability and statistics. As such, understanding of these concepts is necessary for understanding the methods presented in Chapter 4. Readers with little or no background in statistics are encouraged to review chapters 1-5 of the textbook “Statistical Methods in the Atmospheric Sciences” [Wilks, 2019]. Of critical importance are the concepts of a probability distribution function (PDF) and cumulative distribution function (CDF), quantiles of a distribution, extreme values, and frequentist hypothesis testing (e.g. the Student’s \(t\)-test and others). Additional statistical concepts will be introduced in the Guidebook as needed.