Jul 30, 2021 — Baylor University February 15, 2018

Jun 18, 2021 — The final lecture in The D'Arcy Lectures 2021 series, delivered by Revd Dr Patrick Riordan, Senior Fellow in Political Philosophy and Catholic Social Thought. Moderator: Revd Dr Nick Austin SJ, Master of Campion Hall.

Oct 21, 2021 — As the world’s leaders prepare to gather in Glasgow for the UN Climate Change Conference, it is worth reflecting on how each of us as individuals might respond to the environmental crisis. To this end, I propose going back in time to ancient Athens when philosophy began. In that remarkable “golden age”, a trio of philosophers — Socrates, Plato, and Aristotle — thought hard about how each of us should live in ways that are highly relevant to us now.

Mar 30, 2021 — At 11.30 this morning took place, live streaming from the Press Office of the Holy See, the press conference for the presentation of the volume ...

Jan 1, 2021 — Catholic Social Teaching on Care for Creation and Stewardship of the Earth The Catholic Church has a well-documented tradition of Care for Creation and Stewardship of the Earth. This resource includes elements of Catholic teaching that highlight this tradition. This resource is intended to serve as an introduction on this issue; it is not comprehensive. Audience with Representatives of the Churches and Ecclesial Communities of the Different Religions Pope Francis, March 2013 "The Church is likewise conscious of the responsibility which all of us have for our world, for the whole of creation, which we must love and protect. There is much that we can do to benefit the poor, the needy and those who suffer, and to favor justice, promote reconciliation and build peace." Renewing the Earth: An Invitation to Reflection and Action on Environment in Light of Catholic Social Teaching, 1991 (no. 2) "Our mistreatment of the natural world diminishes our own dignity and sacred- ness, not only because we are destroying resources that future generations of hu- mans need, but because we are engaging in actions that contradict what it means to be human. Our tradition calls us to protect the life and dignity of the human person, and it is increasingly clear that this task cannot be separated from the care and de- fense of all creation." World Environment Day, Pope Francis, June 2013 "We are losing the attitude of wonder, contemplation, listening to creation. The implications of living in a horizontal manner [is that] we have moved away from God, we no longer read His signs." DSTARENCE CATHOLIC Renewing the Earth: An Invitation to Reflection and Action on Environment in Light of Catholic Social Teaching, 1991 (no. 8) "Created things belong not to the few, but to the entire human family.” BISHOPS The Compendium of the Social Doctrine of the Church Pontifical Council for Justice and Peace, 2005 (no. 466) "Care for the environment represents a challenge for all of hu- manity. It is a matter of a common and universal duty, that of respecting a common good, destined for all, by preventing any- one from using 'with impunity the different categories of beings, whether living or inanimate—animals, plants, the natural ele- ments simply as one wishes, according to one's own economic needs.' It is a responsibility that must mature on the basis of the global dimension of the present ecological crisis and the conse- quent necessity to meet it on a worldwide level, since all beings are interdependent in the universal order established by the Crea- tor. ‘One must take into account the nature of each being and of its mutual connection in an ordered system, which is precisely the 'cosmos". World Day of Peace, Pope Emeritus Benedict XVI, 2007 "Alongside the ecology of nature, there exists what can be called a 'human' ecology, which in turn demands a 'social' ecology. All this means that humani- ty, if it truly desires peace, must be increasingly conscious of the links between natural ecology, or respect for nature, and human ecology. Experience shows that disregard for the environment always harms human coexistence, and vice versa. It becomes more and more evident that there is an inseparable link be- tween peace with creation and peace among men. Department of Justice, Peace and Human Development 3211 4th St. NE Washington, DC 20017 (202)541-3160 usccb.org/jphd • The Compendium of the Social Doctrine of the Church Pontifical Council for Justice and Peace, 2005 "There is a need to break with the logic of mere consumption and promote forms of agricultural and industrial production that respect the order of creation and satisfy the basic human needs of all. These attitudes, sustained by a renewed awareness of the interde- pendence of all the inhabitants of the earth, will contribute to elimi- nating the numerous causes of ecological disasters as well as guaranteeing the ability to re- spond quickly when such disas- ters strike people and territories. The ecological question must not be faced solely because of the frightening prospects that envi- ronmental destruction represents: rather it must above all become a strong motivation for an authen- tic solidarity of worldwide di- mensions" (no. 486). The Compendium of the Social Doctrine of the Church Pontifical Council for Justice and Peace, 2005 (no. 462) Charity in Truth (Caritas in Veritate) Pope Emeritus Benedict XVI, 2009 Address to Diplomatic Corps, January 2010 Pope Emeritus Benedict XVI "The protection of the environment, of resources and of the climate obliges all international leaders to act justly and to show a readiness to work in good faith, respecting the law and promoting solidarity with the weakest regions of the planet." (no. 50) "With the progress of science and technology, questions as to their meaning increase and give rise to an ever greater need to respect the transcendent dimension of the hu- man person and creation itself." CATHOLIC "[T]his concern and commitment for the environment should be situat- ed within the larger framework of the great challenges now facing man- kind. If we wish to build true peace, how can we separate, or even set at odds, the protection of the environment and the protection of human life, including the life of the unborn? It is in man's respect for himself that his sense of responsibility for creation is shown." On the Development of Peoples (Populorum Progressio), Pope Paul VI, 1967 "Already on the first page of Sacred Scripture we read these words: Fill the earth and subdue (Gn 1:28). By these words we are taught that all things of the world have been created for man, and that this task has been entrust- ed to him to enhance their value by the resources of his intellect, and by his toil to complete and perfect them for his own use. Now if the earth has been created for the purpose of furnishing individuals either with the ne- cessities of a livelihood or the means for progress, it follows that each man has the right to get from it what is necessary for him. The Second Ecumen- ical Vatican Council has reminded us of this in these words: 'God destined the earth with all that it contains for the use of all men and nations, in such a way that created things in fair share should accrue to all men under the leadership of justice with charity as a companion.” (no. 22) BISHOPS "How can we forget, for that matter the struggle for access to natural resources is one of the causes of a number of conflicts, not the least in Africa, as well as a continuing threat elsewhere? For this reason too, I forcefully repeat that to cultivate peace, one must protect creation!" Economic Justice for All, 1997 (no. 34) citing St. Cyprian "From the patristic period to the present, the Church has affirmed that misuse of the world's resources or appropriation of them by a mi- nority of the world's population be- trays the gift of creation since 'whatever belongs to God belongs to all." Global Climate Change: A Plea for Dialogue, Prudence and the Common Good, 2001 "At its core, global climate change is not about economic theory or political platforms, nor about partisan advantage or interest group pressures. It is about the future of God's creation and the one human family. It is about protecting both 'the human environ- ment' and the natural envi- ronment. It is about our hu- man stewardship of God's creation and our responsibil- ity to those who come after us." Department of Justice, Peace and Human Development 3211 4th St. NE .Washington, DC 20017 (202)541-3160 usccb.org/jphd

Dec 16, 2020 — For more than a century, the teaching authority of the Catholic Church has attempted to walk along with the modern world, criticizing what is bad and praising what is good. Counsels of Imperfection described the current state of that fairly bumpy journey. The book is divided into 11 chapters. First comes an introduction to ever-changing modernity and the unchanging Christian understanding of human nature and society. Then come two chapters on economics, including a careful delineation of the Catholic response, past and present, to socialism and capitalism. The next topic is government, with one chapter on Church and State, another on War, and a third that runs quickly through democracy, human rights, the welfare state, crimes and punishments (including the death penalty), anti-Semitism, and migration. Counsels of Imperfection then dedicates two chapters on ecology, including an enthusiastic analysis of Francis’s “technocratic paradigm”. The last topic is the family teaching, which presents the social aspects of the Church’s sexual teaching. A brief concluding chapter looks at the teaching’s changing response to the modern world, and at the ambiguous Catholic appreciation of the modern idea of progress. For each topic, Counsels of Imperfection provides biblical, historical and a broad philosophical background. Thomas Aquinas appears often, but so does G. W. F Hegel. The goal is not only to explain what the Church really says, but also how it got to its current position and who it is arguing with. In the spirit of a doctrine that is always in development, Counsels of Imperfection points out both strong-points and imperfections in the teaching. The book should be of interest to specialists in Catholic Social Teaching, but its main audience is curious newcomers, especially people who do not want to be told that there are simple Catholic answers to the complicated problems of the modern world.

May 1, 2020 — Work is the vocation of man: Morning Meditation by Pope Francis in the Chapel of the Domus Sanctae Marthae, 1st May 2020

Feb 2, 2020 — “Querida Amazonia”: Post-Synodal Exhortation to the People of God and to All Persons of Good Will (2 February 2020)

Jan 1, 2020 — The Philosophy of Climate Science Climate change is one of the defining challenges of the 21st century. But what is climate change, how do we know about it, and how should we react to it? This article summarizes the main conceptual issues and questions in the foundations of climate science, as well as of the parts of decision theory and economics that have been brought to bear on issues of climate in the wake of public discussions about an appropriate reaction to climate change. We begin with a discussion of how to define climate. Even though “climate” and “climate change” have become ubiquitous terms, both in the popular media and in academic discourse, the correct definitions of both notions are hotly debated topics. We review different approaches and discuss their pros and cons. Climate models play an important role in many parts of climate science. We introduce different kinds of climate models and discuss their uses in detection and attribution, roughly the tasks of establishing that the climate of the Earth has changed and of identifying specific factors that cause these changes. The use of models in the study of climate change raises the question of how well-confirmed these models are and of what their predictive capabilities are. All this is subject to considerable debate, and we discuss a number of different positions. A recurring theme in discussions about climate models is uncertainty. But what is uncertainty and what kinds of uncertainties are there? We discuss different attempts to classify uncertainty and to pinpoint their sources. After these science-oriented topics, we turn to decision theory. Climate change raises difficult questions such as: What is the appropriate reaction to climate change? How much should we mitigate? To what extent should we adapt? What form should adaptation take? We discuss the framing of climate decision problems and then offer an examination of alternative decision rules in the context of climate decisions. Table of Contents Introduction Defining Climate and Climate Change Climate Models Detection and Attribution of Climate Change Confirmation and Predictive Power Understanding and Quantifying Uncertainty Conceptualising Decisions Under Uncertainty Managing Uncertainty Conclusion Glossary References and Further Reading 1. Introduction Climate science is an umbrella term referring to scientific disciplines studying aspects of the Earth’s climate. It includes, among others, parts of atmospheric science, oceanography, and glaciology. In the wake of public discussions about an appropriate reaction to climate change, parts of decision theory and economics have also been brought to bear on issues of climate. Contributions from these disciplines that can be considered part of the application of climate science fall under the scope of this article. At the heart of the philosophy of climate science lies a reflection on the methodology used to reach various conclusions about how the climate may evolve and what we should do about it. The philosophy of climate science is a new sub-discipline of the philosophy of science that began to crystalize at the turn of the 21st century when philosophers of science started having a closer look at methods used in climate modelling. It comprises a reflection on almost all aspects of climate science, including observation and data, methods of detection and attribution, model ensembles, and decision-making under uncertainty. Since the devil is always in the detail, the philosophy of climate science operates in close contact with science itself and pays careful attention to the scientific details. For this reason, there is no clear separation between climate science and the philosophy thereof, and conferences in the field are often attended by both scientists and philosophers. This article summarizes the main problems and questions in the foundations of climate science. Section 2 presents the problem of defining climate. Section 3 introduces climate models. Section 4 discusses the problem of detecting and attributing climate change. Section 5 examines the confirmation of climate models and the limits of predictability. Section 6 reviews classifications of uncertainty and the use of model ensembles. Section 7 turns to decision theory and discusses the framing of climate decision problems. Section 8 introduces alternative decision rules. Section 9 offers a few conclusions. Two qualifications are in order. First, we review issues and questions that arise in connection with climate science from a philosophy of science perspective, and with special focus on epistemological and decision-theoretic problems. Needless to say, this is not the only perspective. Much can be said about climate science from other points of view, most notably science studies, sociology of science, political theory, and ethics. For want of space, we cannot review contributions from these fields. Second, to guard against possible misunderstandings, it ought to be pointed out that engaging in a critical philosophical reflection on the aims and methods of climate science is in no way tantamount to adopting a position known as climate scepticism. Climate sceptics are a heterogeneous group of people who do not accept the results of ‘mainstream’ climate science, encompassing a broad spectrum from those who flat out deny the basic physics of the greenhouse effect (and the influence of human activities on the world’s climate) to a small minority who actively engage in scientific research and debate and reach conclusions at the lowest end of climate impacts. Critical philosophy of science is not the handmaiden of climate scepticism; nor are philosophers ipso facto climate sceptics. So, it should be stressed here that we do not endorse climate scepticism. We aim to understand how climate science works, reflect on its methods, and understand the questions that it raises. 2. Defining Climate and Climate Change Climate talk is ubiquitous in the popular media as well as in academic discourse, and climate change has become a familiar topic. This veils the fact that climate is a complex concept and that the correct definitions of climate and climate change are a matter of controversy. To gain an understanding of the notion of climate, it is important to distinguish it from weather. Intuitively speaking, the weather at a particular place and a particular time is the state of the atmosphere at that place and at the given time. For instance, the weather in central London at 2 pm on 1 January 2015 can be characterised by saying that the temperature is 12 degrees centigrade, the humidity is 65%, and so forth. By contrast, climate is an aggregate of weather conditions: it is a distribution of particular variables (called the climate variables) arising for a particular configuration of the climate system. The question is how to make this basic idea precise, and this is where different approaches diverge. 21st-century approaches to defining climate can be divided into two groups: those that define climate as a distribution over time, and those that define climate as an ensemble distribution. The climate variables in both approaches include those that describe the state of the atmosphere and the ocean, and sometimes also variables describing the state of glaciers and ice sheets [IPCC 2013]. Distribution over time. The state of the Earth depends on external conditions of the system such as the amount of energy received from the sun and volcanic activity. Assume that there is a period of time over which the external conditions are relatively stable in that they exhibit small fluctuations around a constant mean value c. One can then define the climate over this time period as the distribution of the climate variables over that period under constant external conditions c [for example, Lorenz 1995]. Climate change then amounts to successive time periods being characterised by different distributions. However, in reality the external conditions are not constant and even when there are just slight fluctuations around c, the resulting distributions may be very different. Hence this definition is unsatisfactory [Werndl 2015]. This problem can be avoided by defining climate as the empirically observed distribution over a specific period of time, where external conditions are allowed to vary. Again, climate change amounts to different distributions for successive time periods. This definition is popular because it is easy to estimate from the observations, for example, from the statistics taken over thirty years that are published by the World Meteorological Organisation [Hulme et al. 2009]. A major problem of this definition can be illustrated by the example in which, in the middle of a period of time, the Earth is hit by a meteorite and becomes a much colder place. Clearly, the climate before and after the hit of the meteor differ. Yet this definition has no resources to recognize this because all it says is that climate is a distribution arising over a specific time period. To circumvent this problem, Werndl [2015] introduces the idea of regimes of varying external conditions and suggests defining climate as the distribution over time of the climate variables arising under a specific regime of varying external conditions. The challenge for this account is to spell out what exactly is meant by a regime of varying external conditions. Ensemble Distribution. An ensemble of climate systems (not to be confused with a model ensemble) is an infinite collection of virtual copies of the climate system. Consider the sub-ensemble of these that satisfy the condition that the present values of the climate variables lie in a specific interval around the values measured in the actual climate system (that is, the values compatible with the measurement accuracy). Now assume again that there is period of time over which the external conditions are relatively stable in that they exhibit small fluctuations around a constant mean value c. Then climate at future time t is defined as the distribution of values of the climate variables that arises when all systems in the ensemble evolve from now to t under constant external conditions c [for example, Lorenz 1995]. In other words, the climate in the future is the distribution of the climate variables over all possible climates that are consistent with current observations under the assumption of constant external conditions c. As we have seen previously, in reality, external conditions are not constant and even small fluctuations around a mean value can lead to different distributions [Werndl 2015]. This worry can be addressed by tracing the development of the initial condition ensemble under actual external conditions. The climate at future time t then is the distribution of the climate variables that arises when the initial conditions ensemble is evolved forward for the actual path taken by the external conditions [for example, Daron and Stainforth 2013]. This definition faces a number of conceptual challenges. First, it makes the world’s climate dependent on our knowledge (via measurement accuracy), but this is counterintuitive because we think of climate as something objective that is independent of our knowledge. Second, the above definition is a definition of future climate, and it is difficult to see how the present and past climate should be defined. Yet without a notion of the present and past climate one cannot define climate change. A third problem is that ensemble distributions (and thus climate) do not relate in a straightforward way to the past time series of observations of the actual Earth and this would imply that the climate cannot be estimated from them [compare, Werndl 2015]. These considerations show that defining climate is nontrivial and there is no generally accepted or uncontroversial definition of climate. 3. Climate Models A climate model is a representation of particular aspects of the climate system. One of the simplest climate models is an energy-balance model, which treats the Earth as a flat surface with one layer of atmosphere above it. It is based on the simple principle that in equilibrium the incoming and outgoing radiation must be equal (see Dessler [2011], Chapters 3-6, for a discussion of such models). This model can be refined by dividing the Earth into zones, allowing energy transfer between zones, or describing a vertical profile of the atmospheric characteristics. Despite their simplicity, these models provide a good qualitative understanding of the greenhouse effect. Modern climate science aims to construct models that integrate as much as possible of the known science (for an introduction to climate modelling see [McGuffie and Henderson-Sellers 2005]). Typically, this is done by dividing the Earth (both the atmosphere and ocean) into grid cells. In 2020, global climate models have a horizontal grid scale of around 150 km. Climatic processes can then be conceptualised as flows of physical quantities such as heat or vapour from one cell to another. These flows are mathematically described by equations. These equations form the ‘dynamical core’ of a global circulation model (GCM). The equations typically are intractable with analytical methods, and powerful supercomputers are used to solve them. For this reason, they are often referred to as ‘simulation models’. To solve equations numerically, time is discretised. Current state-of-the-art simulations use time steps of approximately 30 minutes, taking weeks or months in real time on supercomputers to simulate a century of climate evolution. In order to compute a single hypothetical evolution of the climate system (a ‘model run’), we also require an initial condition and boundary conditions. The former is a mathematical description of the state of the climate system (projected into the model’s own domain) at the beginning of the period being simulated. The latter are values for any variables which affect the system, but which are not directly output by the calculations. These include, for instance, the concentration of greenhouse gases, the amount of aerosols in the atmosphere at a given time, and the amount of solar radiation received by the Earth. Since these are drivers of climatic change, they are often referred to as external forcings or external conditions. Where processes occur on a smaller scale than the grid, they may be included via parameterisation, whereby the net effect of the process is separately calculated as a function of the grid variables. For instance, cloud formation is a physical process that cannot be directly simulated because typical clouds are much smaller than the grid. So, the net effect of clouds is usually parameterised (as a function of temperature, humidity, and so forth) in each grid cell and fed back into the calculation. Sub-grid processes are one of the main sources of uncertainty in climate models. There are now dozens of global climate models under continuous development by national modelling centres like NASA, the UK Met Office, and the Beijing Climate Center, as well as by smaller institutions. An exact count is difficult because many modelling centres maintain multiple versions based on the same foundation.  As an indication, in 2020 there were 89 model-versions submitted to CMIP6 (Coupled Model Intercomparison Project phase 6), from 35 modelling groups, though not all of these should be thought of as being “independent” models since assumptions and algorithms are often shared between institutions. In order to be able to compare outputs of these different models, the Coupled Model Intercomparison Project (CMIP) defines a suite of standard experiments to be run for each climate model. One standard experiment is to run each model using the historical forcings experienced during the twentieth century so that the output can be directly compared against real climate system data. Climate models are used in many places in climate science, and their use gives rise to important questions. These questions are discussed in the next three sections. 4. Detection and Attribution of Climate Change Every empirical study of climate has to begin by observing the climate. Meteorological observatories measure a number of variables such as air temperature near the surface of the Earth using thermometers. But more or less systematic observations are available since about 1750, and hence to reconstruct the climate before then scientists have to rely on proxy data: data for climate variables that are derived from observing other natural phenomena such as tree rings, ice cores, and ocean sediments. The use of proxy data raises a number of methodological problems centred around the statistical processing of such data, which are often sparse, highly uncertain, and several inferential steps away from the climate variable of interest. These issues were at the heart of what has become known as the Hockey Stick controversy, which broke at the turn of the century in connection with a proxy-based reconstruction of the Northern Hemisphere temperature record [Mann, Bradley and Hughes, 1998]. The sceptics pursued two lines of argument. They cast doubt on the reliability of the available data, and they argued that the methods used to process the data are such that they would produce a hockey-stick-shaped curve from almost any data [for example, McIntyre and McKitrick 2003]. The papers published by the sceptics raised important issues and stimulated further research, but they were found to contain serious flaws undermining their conclusions. There are now more than two dozen reconstructions of this temperature record using various statistical methods and proxy data sources. Although there is indeed a wide range of plausible past temperatures, due to the constraints of the data and methods, these studies do robustly support the consensus that, over the past 1400 years, temperatures during the late 20th century are likely to have been the warmest [Frank et al. 2010]. Do rising temperatures indicate that there is climate change, and if so, can the change be attributed to human action? These two problems are known as the problems of detection and attribution. The Intergovernmental Panel on Climate Change (IPCC) defines these as follows: Detection of change is defined as the process of demonstrating that climate or a system affected by climate has changed in some defined statistical sense without providing a reason for that change. An identified change is detected in observations if its likelihood of occurrence by chance due to internal variability alone is determined to be small […]. Attribution is defined as ‘the process of evaluating the relative contributions of multiple causal factors to a change or event with an assignment of statistical confidence.’ [IPCC 2013] These definitions raise a host of issues. The root cause of the difficulties is the clause that climate change has been detected only if an observed change in the climate is unlikely to be due to internal variability. Internal variability is the phenomenon that the values of climate variables such as temperature and precipitation would change over time due to the internal dynamics of the climate system even in the absence of a change in external conditions, because of fluctuations in the frequency of storms, ocean currents, and so on. Taken at face value, this definition of detection has the consequence that there cannot be internal climate change. The ice ages, for instance, would not count as climate change if they occurred because of internal variability. This is not only at odds with basic intuitions about climate and with the most common definitions of climate as a finite distribution over a relatively short time period (where internal climate change is possible); it also leads to difficulties with attribution: if detected climate change is ipso facto change not due to internal variability, then from the very beginning it is excluded that particular factors (namely, internal climate dynamics) can lead to a change in the climate, which seems to be an unfortunate conclusion. For the case of the ice ages, many researchers would stress that internal variability is different from natural variability. Since, say, orbital changes explain the ice ages, and orbital changes are natural but external, this is a case of external climate change. While this move solves some of the problems, there remains the problem that there is no generally accepted way to separate internal and external factors, and the same factor is sometimes classified as internal and sometimes as external. For instance, glaciation processes are sometimes treated as internal factors and sometimes as prescribed external factors. Likewise, sometimes the biosphere is treated as an external factor, but sometimes it is also internally modelled and treated as an internal factor. One could even go so far to ask whether human activity is an external forcing on the climate system or an internally-generated Earth system process. Research studies usually treat human activity as an external forcing, but it could consistently be argued that human activities are an internal dynamical process. The appropriate definition simply depends on the research question of interest. For a discussion of these issues, see Katzav and Parker [2018]. The effects of internal variability are present on all timescales, from the sub-daily fluctuations experienced as weather to the long-term changes due to cycles of glaciation. Since internal variability stems from processes in a highly complex nonlinear system, it is also unlikely that the statistical properties of internal variability are constant over time, which further compounds methodological difficulties. State-of-the-art climate models run with constant forcing show significant disagreements both on the magnitude of internal variability and the timescale of variations. (On http://www.climate-lab-book.ac.uk/2013/variable-variability/#more-1321 the reader finds a plot showing the internal variability of all CMIP5 models. The plot indicates that models exhibit significantly different internal variability, leaving considerable uncertainty.) The model must be deemed to simulate pre-industrial climate (including variability) sufficiently well before it can be used for such detection and attribution studies, but we do not have thousands of years of detailed observations upon which to base that judgement. Estimates of internal variability in the climate system are produced from climate models themselves [Hegerl et al. 2010], leading to potential circularity. This underscores the difficulties in making attribution statements based on the above definition, which recognises an observed change as climate change only if is unlikely to be due to internal variability. Since the IPCC’s definitions are widely used by climate scientists, the discussion about detection and attribution in the remainder of this section is based on these definitions. Detection relies on statistical tests, and detection studies are often phrased in terms of the likelihood of a particular event or sequence of events happening in the absence of climate change. In practice, the challenge is to define an appropriate null hypothesis (the expected behaviour of the system in the absence of changing external influences), against which the observed outcomes can be tested. Because the climate system is a dynamical system with processes and feedbacks operating on all scales, this is a non-trivial exercise. An indication of the importance of the null hypothesis is given by the results of Cohn and Lins [2005], who compare the same data against alternate null hypotheses, with results differing by 25 orders of magnitude of significance! This does not in itself show that either null is more appropriate, but it demonstrates the sensitivity of the result to the null hypothesis chosen. This, in turn, underscores the importance of the choice of null hypothesis and the difficulty of making any such choice if the underlying processes are poorly understood. In practice, the best available null hypothesis is often the best available model of the behaviour of the climate system, including internal variability, which for most climate variables usually means a state of the art GCM. This model is then used to perform long control runs with constant forcings in order to quantify the internal variability of the model (see discussion above). Climate change is then said to have been detected when the observed values fall outside a predefined range of the internal variability of the model. The difficulty with this method is that there is no single “best” model to choose: many such models exist, they are similarly well developed, but, as noted above, they have appreciably different patterns of internal variability. The differences between different models are relatively unimportant for the clearest detection results such as recent increases in global mean temperature. Here, as stressed by Parker [2010], detection is robust across different models (for a discussion of robustness see Section 6), and, moreover, there is a variety of different pieces of evidence all pointing to the conclusion that the global mean temperature has increased beyond that which can be attributed to internal variability. However, the issues of which null hypothesis to use and how to quantify internal variability, can be important for the detection of subtler local climate change. If climate change has been detected, then the question of attribution arises. This might be an attribution of any particular change (either a direct climatic change such as increased global mean temperature, or an impact such as the area burnt by forest fires) to any identified cause (such as increased CO2 in the atmosphere, volcanic eruptions, or human population density). Where an impact is considered, a two-step or multi-step approach may be appropriate. An example of this, taken from the IPCC Good Practice Guidance paper [Hegerl et al. 2010], is the attribution of coral reef calcification impacts to rising CO2 levels, in which an intermediate stage is used by first attributing changes in the carbonate ion concentration to rising CO2 levels, then attributing calcification processes to changes in the carbonate ion concentration. This also illustrates the need for a clear understanding of the physical mechanisms involved, in order to perform a reliable multi-step attribution in the presence of many potential confounding factors. In the interpretation of attribution results, in particular those framed as a question of whether human activity has influenced a particular climatic change or event, there is a tendency to focus on whether the confidence interval of the estimated anthropogenic effect crosses zero. The absence of such a crossing indicates that change is likely to be due to human factors. This results in conservative attribution statements, but it reflects the focus of the present debate where, in the eyes of the public and media, “attribution” is often understood as confidence in ruling out non-human factors, rather than as giving a best estimate or relative contributions of different factors. Statistical analysis quantifies the strength of the relationship, given the simplifying assumptions of the attribution framework, but the level of confidence in the simplifying assumptions must be assessed outside that framework. This level of confidence is standardised by the IPCC into discrete (though subjective) categories (“very high”, “high”, and so forth.), which aim to take account of the process knowledge, data limitations, adequacy of models used, and the presence of potential confounding factors. The conclusion that is reached will then have a form similar to the IPCC’s headline attribution statement: It is extremely likely [³95% probability] that more than half of the observed increase in global average surface temperature from 1951 to 2010 was caused by the anthropogenic increase in greenhouse gas concentrations and other anthropogenic forcings together. [IPCC 2013; Summary for Policymakers, section D.3]. One attribution method is optimal fingerprinting. The method seeks to define a spatio-temporal pattern of change (fingerprint) associated with each potential driver (such as the effect of greenhouse gases or of changes in solar radiation), normalised relative to the internal variability, and then perform a statistical regression of observed data with respect to linear combinations of these patterns. The residual variability after observations have been attributed to each factor should then be consistent with the internal variability; if not, this suggests that an important source of variability remains unaccounted for. Parker [2010] notes that fingerprint studies rely on several assumptions. Chief among them is linearity, that is, that the response of the climate system when several forcing factors are present is equal to a linear combination of the effects of the forcings. Because the climate system is nonlinear, this is clearly a source of methodological difficulty, although for global-scale responses (in contrast to regional-scale responses) additivity has been shown to be a good approximation. Levels of confidence in these attribution statements are primarily dependent on physical understanding of the processes involved.  Where there is a clear, simple, well-understood mechanism, there should be greater confidence in the statistical result; where the mechanisms are loose, multi-factored or multi-step, or where a complex model is used as an intermediary, confidence is correspondingly lower.  The Guidance Paper cautions that, Where models are used in attribution, a model’s ability to properly represent the relevant causal link should be assessed. This should include an assessment of model biases and the model’s ability to capture the relevant processes and scales of interest. [Hegerl 2010, 5] As Parker [2010] argues, there is also higher confidence in attribution results when the results are robust and there is a variety of evidence. For instance, the finding that late twentieth-century temperature increase was mainly caused by greenhouse gas forcing is found to be robust given a wide range of different models, different analysis techniques, and different forcings; and there is a variety of evidence all of which supports this claim. Thus our confidence that greenhouse gases explain global warming is high. (For further useful extended discussion of detection and attribution methods in climate science, see pages 872-878 of IPCC [2013] and in the Good Practice Guidance paper by Hegerl et al. [2010], and for a discussion of how such hypotheses are tested see Katzav [2013].) In addition to the large-scale attribution of climate change, attribution of the degree to which individual weather events have become either more likely or more extreme as a result of increasing atmospheric greenhouse gas concentrations is now common. It has a particular public interest as it is perceived as a way both to communicate that climate impacts are happening already, perhaps quantifying risk numerically to price insurance, and offering a motivation for climate mitigation.  There is therefore also an incentive to conduct these studies quickly, to inform timely news articles, and some groups have formed to respond quickly to reports of extreme weather and conduct attribution studies immediately. This relies on the availability of data, may suffer from unclear definitions of exactly what category of event is being analysed, and is open to criticism for publicity prior to peer review.  There are also statistical implications of choosing to analyse only those events which have happened and not those that did not happen. For a discussion of event attribution see Lloyd and Oreskes [2019] and Lusk [2017]. 5. Confirmation and Predictive Power Two questions arise in connection with models: how are models confirmed and what is their predictive power? Confirmation concerns the question of whether, and to what degree, a specific model is supported by the data. Lloyd [2009] argues that many climate models are confirmed by past data. Parker [2009] objects to this claim. She argues that the idea that climate models per se are confirmed cannot be seriously entertained because all climate models are known to be wrong and empirically inadequate. Parker urges a shift in thinking from confirmation to adequacy for purpose: models can only be found to be adequate for specific purposes, but they cannot be confirmed wholesale. For example, one might claim that a particular climate model adequately predicts the global temperature increase that will occur by 2100 (when run from particular initial conditions and relative to a particular emission scenario). Yet, at the same time, one might hold that the predictions of global mean precipitation by 2100 by the same model cannot be trusted. Katzav [2014] cautions that adequacy for purpose assessments are of limited use. He claims that these assessments are typically unachievable because it is far from clear which of the model’s observable implications can possibly be used to show that the model is adequate for the purpose. Instead, he argues that climate models can at best be confirmed as providing a range of possible futures. Katzav is right to stress that adequacy for purpose assessments are more difficult than appears at first sight. But the methodology of adequacy for purpose cannot be dismissed wholesale; in fact, it is used successfully across the sciences (for example, when ideal gas models are confirmed to be useful for particular purposes). Whether or not adequacy for purpose assessment is possible depends on the case at hand. If one finds that one model predicts specific variables well and another model doesn’t, then one would like to know the reasons why the first model is successful and the second not. Lenhard and Winsberg [2010] argue that this is often very difficult, if not impossible: For complex climate models a strong version of confirmation holism makes it impossible to tell where the failures and successes of climate models lie. In particular, they claim that it is impossible to assess the merits and problems of sub-models and the parts of models. There is a question, though, whether this confirmation holism affects all models and whether it is here to stay. Complex models have different modules for the atmosphere, the ocean, and ice. These modules can be run individually and also together. The aim of the many new Model Intercomparison Projects (MIPs) is, by comparing individual and combined runs, to obtain an understanding of the performance and physical merits of separate modules, which it is hoped will identify areas for improvement and eventually result in better performance of the entire model. Another problem concerns the use of data in the construction of models. The values of model parameters are often estimated using observations, a process known as calibration. For example, the magnitude of the aerosol forcing is sometimes estimated from data. When data have been used for calibration, the question arises whether the same data can be used again to confirm the model. If data are used for confirmation that have not already been used for calibration, they are use-novel. If data are used for both calibration and confirmation, this is referred to as double-counting. Scientists and philosophers alike have argued that double-counting is illegitimate and that data have to be use-novel to be confirmatory [Lloyd 2010; Shackley et al. 1998; Worrall 2010]. Steele and Werndl [2013] oppose this conclusion and argue that on Bayesian and relative-likelihood accounts of confirmation double-counting is legitimate. Furthermore, Steele and Werndl [2015] argue that model selection theory presents a more nuanced picture of the use of data than the commonly endorsed positions. Frisch [2015] cautions that Bayesian as well as other inductive logics can be applied in better and worse ways to real problems such as climate prediction. Nothing in the logic prevents facts from being misinterpreted and their confirmatory power exaggerated (as in ‘the problem of old evidence’ which Frisch [2015] discusses). This is certainly a point worth emphasising. Indeed, Steele and Werndl [2013] stress that the same data cannot inform a prior probability for a hypothesis and also further (dis)confirm the hypothesis. But they do not address all the potential pitfalls in applying Bayesian or other logics to the climate and other settings. Their argument must be understood as a limited one: there is no univocal logical prohibition against the same data serving for calibration and confirmation. As far as non-Bayesian methods of model selection goes, there are two cases. First, there are methods such as cross-validation where the data are required to be use-novel. For cross-validation, the data are split up into two groups: the first group is used for calibration and the second for confirmation. Second, there are the methods such as the Akaike Information Criterion for which the data need not be use-novel, although information criteria methods are hard to apply in practice to climate models because the number of degrees of freedom is poorly defined. This brings us to the second issue: prediction. In the climate context this is typically framed as the issue of projection. ‘Projection’ is a technical term in the climate modelling literature and refers to a prediction that is conditional on a particular forcing scenario and a particular initial conditions ensemble. The forcing scenario is specified either by the amount of greenhouse gas emissions and aerosols added to the atmosphere or directly by their atmospheric concentrations, and these in turn depend on future socioeconomic and technological developments. Much research these days is undertaken with the aim of generating projections about the actual future evolution of the Earth system under a particular emission scenario, upon which policies are made and real-life decisions are taken. In these cases, it is necessary to quantify and understand how good those projections are likely to be. It is doubtful that this question can be answered along traditional lines. One such line would be to refer to the confirmation of a model against historical data (Chapter 9 of IPCC [2013] discusses model evaluation in detail) and argue that the ability of a model to successfully reproduce historical data should give us confidence that it will perform well in the future too. It is unclear at best whether this is a viable answer. The problem is that climate projections for high forcing scenarios take the system well outside any previously experienced state, and at least prima facie there is no reason to assume that success in low forcing contexts is a guide to success in high-forcing contexts; for example, a model calibrated on data from a world with the Arctic Sea covered in ice might no longer perform well when the sea ice is completely melted and the relevant dynamical processes are quite different. For this reason, calibration to past data has at most limited relevance for the assessment of a model’s predictive success [Oreskes et al. 1994; Stainforth et al. 2007a, 2007b, Steele and Werndl 2013]. This brings into focus the fact that there is no general answer to the question of the trustworthiness of model outputs. There is widespread consensus that predictions are better for longer time averages, larger spatial averages, low specificity and better physical understanding; and, all other things being equal, shorter lead times (nearer prediction horizons) are easier to predict than longer ones. Global mean temperature trends are considered trustworthy, and it is generally accepted that the observed upward trend will continue [Oreskes 2007], although the basis of this confidence is usually a physical understanding of the greenhouse effect with which the models are consistent, rather than a direct reliance on the output of models themselves. A 2013 IPCC report [IPCC 2013, Summary for Policymakers, section D.1] professes that modelled surface temperature patterns and trends are trustworthy on the global and continental scale, but, even in making this statement, assigns a probability of at least 66% (‘likely’) to the range within with 90% of model outcomes fall. In plainer terms, this is an expert-assigned probability of at least tens of percent that the models are substantially wrong even

Dec 1, 2019 — Message sent by the Holy Father Francis to Ms. Carolina Schmidt, Minister of the Environment of Chile and President of Cop 25, and to the participants in the United Nations Conference on Climate, 1 December 2019