The Lorentz Principles for the Communication of Model-based Information

Leonard A Smith, Arthur C Petersen, and Erica L Thompson

Statement of Principles PDF

WHY?
For a given decision, some information will be relevant and adequate for the purpose of informing that decision. A single forecast may be adequate for some purposes and not others. For instance, if today is Wednesday and I wish to decide whether to hold a party this weekend on Saturday or Sunday, the weather forecast will be a useful input. But if I am setting a date for a party in three months’ time, and wish to decide between the Saturday and Sunday of a given weekend, then even today’s best available weather forecast is not adequate for this purpose. Decision-makers would benefit if it were made clearer where today’s best available information is thought to be, or not be, adequate for different kinds of decision support.

TELL ME MORE 
How can we assess adequacy for a given purpose?  In the example of a weather forecast, this is reasonably easy: we can look at a large sample of previous weather forecasts for a similar situation (location, lead time, type of weather) and, with reference to observations of the actual weather, generate a statistical description of the forecast quality. This can then to be used to decide, for each purpose, whether this weather forecast is adequate as a source of information, or not.

This relies on the existence of a relevant archive of previous forecasts and outcomes. Where we are forecasting climate, firstly we may have fewer or no previous forecasts to assess, and secondly, we are not sure whether skill in the past will be directly related to skill in the future, since the underlying climate is itself known to be changing.

References and further reading

Parker, W. S. (2009, June). Confirmation and adequacy‐for‐purpose in climate modelling. In Aristotelian Society Supplementary Volume (Vol. 83, No. 1, pp. 233-249). Oxford, UK.

Smith, L.A. (2002) ‘What might we learn from climate forecasts?‘, Proc. National Acad. Sci. USA, 4 99: 2487-2492. DOI: 10.1073/pnas.012580599. Abstract.

Comments on Principle 1
Reason Machete – P01-0719
The example given in the WHY is a weather example. I think it would be better to contrast two climate scenarios. For instance, information required to decide which crop to farm next season in a given region compared with deciding whether to build a dam today based on decadal projections. Of course the weather example illustrates the point. 

Dewi Le Bars – P01-0725
The weather forecast example is a nice illustration. Unfortunately, in practice in climate science most of the time there is no clear cut between information that is adequate or not adequate for purpose. Making that decision is difficult and requires a dialogue between “expert users” and climate scientists. An example for sea level: Hinkel, J., Church, J. A., Gregory, J. M., Lambert, E., Le Cozannet, G., Lowe, J., … Wal, R. (2019). Meeting User Needs for Sea Level Rise Information: A Decision Analysis Perspective. Earth’s Future, 7(3), 320–337. https://doi.org/10.1029/2018EF001071 

WHY?
A good decision is one where, with hindsight, we still feel happy about the process used to arrive at the decision. This does not necessarily mean having made “the right decision” in every circumstance: for example, if we are choosing whether to act given a forecast which is probabilistic. The “limits of today’s science” are the possibilities of quantifying these outcomes rigorously, where the future is not perfectly known.

Without this knowledge, it is harder to make good decisions. If we expect to be able to rely on an hour-by-hour weather forecast for next month, we may be disappointed when the outcomes are not as we expected. In hindsight, having evaluated the forecast, we realise that the procedure was not robust due to an unreliable input. With better communication about the limitations of the science, we could have made a better (more robust) decision.

TELL ME MORE
It may be the right decision to choose not to take an umbrella on days when there is a 10% chance of rain forecast by a reliable system; on 10% of those days, then, you will get wet. But you can still be content that the decision-making process is one which optimises your outcome overall. If you have a more conservative utility function and prefer never to get wet, then your threshold for choosing not to take the umbrella could be higher; the point is that you get what you asked for, even if that is getting wet on one in ten rainy days.

Comments on Principle 2
Dewi Le Bars – P02-0725
This is an important principle. However, in practice, in climate science, I see it as an ideal goal that is impossible to reach. How to communicate clearly something that is not clear? Climate projections for the year 2100 cannot be validated. On the other hand confidence in the field of climate science can be built from the understanding of the current earth system dynamics and from past climate projections for the years 2010s. These information do not give a clear limit of the science but an understanding of what can be projected more or less well. The difficulty is then on the users side to decide wether the climate information is robust enough for their decision or not.

WHY?
All models have a range of applicability beyond which it is inadvisable to go. Without quantitative statements about where this range of applicability ends, decision-makers are left with a binary choice between assuming that scientific information is perfect or rejecting it completely. For how long is the model likely to be useful? In what areas is it more or less reliable? Better understanding of the limits of relevance helps decision-makers incorporate scientific information into their decisions.

TELL ME MORE
When we make use of a weather forecast, we understand that tomorrow’s forecast is quite reliable, that next week’s forecast is indicative but liable to change, and that the forecast for a month’s time is usually not worth looking at (although there is value for certain users at this time scale).

Climate forecasts (projections), by contrast, are typically presented to 2100 without an indication of the degree to which this timescale reflects robustness and confidence. An example of good practice, which we recommend to others, is the IPCC AR5 headline projections for 2100 as summarised in WG1 Table SPM.2. This states that the range in which 90% of modelled global mean surface temperature changes fall, for each scenario, is assessed to be a likely (66-100%) range, after accounting for additional uncertainties or different levels of confidence in models. In plainer terms, they expect a chance of up to about one-in-four that the actual global mean surface temperature would fall outside the 90% range of models, even if the scenario were followed exactly.

A similar approach could be taken for quantities derived from these global models, for example the outputs of climate impacts models, downscaling models, and integrated assessment models. In the first instance a simple propagation of the IPCC statement above through to the outputs of the second model would be informative even before the robustness of the second model itself has been quantified in a similar way.

References and further reading

Thompson, E., Frigg, R. and Helgeson, C. (2016) ‘Expert judgment for climate change adaptation‘, Philosopy of Science, 83 5 (December 2016) pp.1110-1121. DOI:10.1086/687942.

H. Tennekes. Protesting against dogma. Energy and Environment, 17(4):609–612, 2006.

Comments on Principle 3
Reason Machete – P03-0724
On the WHY, I find the question of how long the model is likely to be useful a bit unclear. The TELL ME MORE appears to shed light on this question when it talks about lead times of weather and climate forecasts. Does the “how long” refer to lead times of projections (or forecasts)? If that is the case, then I think the question should be “What is the range of lead times for which the model is likely to be useful?” or something along that line. In the statement of the principle, there is mention of “potential irrelevance.” It doesn’t seem obvious to me what potential irrelevance means. Maybe a line in the WHY or TELL ME MORE should be added to explain this point. An example could be helpful. Here is my faint idea of what potential irrelevance means. Climate regional climate models might tell us that Botswana will have dryer summers and wetter winters by 2050. Nonetheless, the information is irrelevant for a blanket water management policy over the entire country because regional variations of climatic and socioeconomic factors have to be taken into account.

Thank you for your comment. I agree that the tell-me-more can be clarified. What we intend is something along the lines of your “ What is the range of lead times for which the model is likely to be useful?”. To take an extreme case of “potential irrelevance”, consider the initial HadSM3 simulations run under climateprediction.net . In some runs a region in the model-Pacific reduced the temperature of the model-planet in an unphysical manner to very low values. Those runs were “irrelevant” in that the model did something that was physically hogwash: one would not want to include those simulations in model-based support for tasks of future planning. Less dramatic examples are common do to assumptions made (today) in all CMIP climate models. We were intending something more focused on the climate model’s shortcomings; aspects of the model mathematical structure that make the model unable to reproduce “the future” even after we observe what the future holds. (LAS)

Hailiang Du – P03-0726
The quantitative statement about future climate based on current climate model is likely to be questionable. But the climate model can still provide insight about “probable” future climate scenarios, which could be valuable to decision makers.

I am not sure what exactly you mean by ‘”probable” future climate scenarios’, CMIP5 models differ significantly regarding the Earth’s global mean temperature, nevertheless each and every of these very different model planets show significant warming over the past 100 years. It seems reasonable to take from this that anthropogenic impacts will warm Earth-like model-planets in general; but I do not see how this can give us quantitative scenarios for the Earth itself without significant additional assumptions regarding the fidelity of the model(s). (LAS)

WHY?
Computational models often use variable names corresponding to the real-world quantities they represent, sometimes even real-world observable/measurable quantities, and the variables in. This is done for obvious reasons as it makes the modelling process more intuitive and explainable. While this simplifies coding the model significantly, it is critical to distinguish (for example) model-temperative when presenting results. Variables in mathematical models correspond to yet another type of entity.

Model Intercomparison Projects (MIPs) can be useful for advancing the art modelling, but model-model comparisons often tell us little about the real world. Indeed some of the most popular CMIP graphs have no connection with reality. Reality Intercomparison Projects might better clarify the level of confidence we place in model simulations by comparing each of them with observations more directly.

For clarity, for good science, and to ensure that the strength and weakness of model-based conclusions are made clear to those using them. It is important not to confuse the properties of a model with the properties of reality.

TELL ME MORE
Example: the measurable viscosity of a physical fluid may not be the same as the “viscosity” variable used in a numerical integration scheme to model that fluid, as the latter may depend on grid spacing and time step. In climate simulations, an “eddy viscosity” or effective viscosity is used, which takes the role of the molecular viscosity but encompasses other sub-grid-scale dissipative processes.

As the numerical value of the eddy viscosity is several orders of magnitude different from the molecular viscosity, the two are always distinguished, but in other cases the physical-variable and model-variable may have such similar values (one observable, the other either derived from physical principles, calibrated with respect to data, or assumed) that they are given the same name.

Properties of model ensembles, such as the “climate sensitivity”, are sometimes assumed to be informative about a real-world climate sensitivity, but the real world is a single system which is not generated in the same way as our class of models. Statistical methods are hamstrung if they assume that the real world is statistically indistinguishable from our class of models.

Simpler cases are worth considering, too. The “altitude above sea level” is a climate model-variable on the scale of tens of kilometres, which averages out sharp peaks such as the Andes, where the “altitude above sea level” in the real world can be more “spiky”, with dynamical consequences for circulation and orographically-induced rainfall. The “freezing point of water” may be set to the measured value (zero Celsius) in a model, but how do we know that we might not get a better model by treating this as a calibration parameter? If this sounds like a crazy idea, consider that perhaps the widespread use of anomaly rather than absolute temperatures is already treating the freezing point of water as a calibration parameter, with results varying by around 3 Celsius (IPCC AR5 WG1, figure 9.8a).

Comments on Principle 4
Luke Bevan – P04-2307
I think that to some extent the overlap in langage can be caused by a difference in epistemological understanding what a model is. There is some research that shows that different people can understand what modelling is in very different ways. For example, some may equate modelling for example with a form of experimentation, some may see them as simplification tools and others may unreflectively consider modelling to be a kind of replication/reconstruction of reality. I guess a useful question is how to communicate the intended relationship between model and either theory, data or other elements that have been used in its construction?

There are indeed relevant, deep and open philosophy of science questions regarding the epistemological status of modelling versus experimentation (or observation) of the real world, on the one hand, and versus theory (or even pure mathematics), on the other hand. The aim of Lorentz Principle 4 is to sensitise modellers to the problem of unreflectively equating a computational model to analytic mathematics and to the real world. Our ‘philosophy of science’ is that making these equal is a (very problematic) assumption that should be avoided. Still, arguments can be made in some cases that models are ‘good enough’ for the purpose to inform real-world decisions and that they do not suffer from spurious results due to computational uncertainties. The point of the principle is to force practitioners to explicitly make their (fallible) arguments. (AP)

Hailiang Du – P04-0726
It would be also useful to clarify under what conditions (if exist), a), b) and c) will share the same/similar properties/behaviours. 

WHY?
By a “traceable account of uncertainty”, we mean a system which allows the reader to trace the path of uncertainty back to its ultimate sources; or conversely, from the (many) sources to the final statement. For example, some uncertainties may be due to measurement imprecision, some may be due to variation between models, and some may be due to inadequacies in models (among other things). The traceable account would identify the chain of uncertainty, showing how each contribution is propagated, quantifying where possible and offering descriptions elsewhere. This would give confidence in the final statement. Scientific uses of the traceable account include identifying the largest sources of uncertainty for further research.

TELL ME MORE
The traceable account is not a new concept, and although easy to describe, it is difficult to implement for large and complex bodies of work. For example, the IPCC offer a traceable account of their conclusions (not limited to uncertainty) by a system of signposting from summaries, to chapters, to journal papers in the literature on which the reports are based.

This can be detrimental to the readability at all levels. It is also an extremely challenging task to retrospectively create a traceable account of uncertainty for a large body of work. For a traceable account of uncertainty to be achievable, it must be built up piece-by-piece, with contributions in a reasonably standardised format by authors of every paper. It is unfeasible, for instance, for authors of a study of climate impacts in West Africa to have to recreate an uncertainty analysis for the global climate models on which those impacts are based, of which they probably have no more than a download of relevant variables. But, those authors must be responsible for cataloguing all of the input and output uncertainties relevant to their own study or model, so that the latter can be used by anyone building on the study.

We have seen this confusion repeatedly from those using the outputs of global climate models (“which one is the best to use as an input for my impact model?” / “how can I decide which to use, if I only have time to do two?” / “so what is the uncertainty in the projection?” / etc). There would be huge community benefit to widespread adoption of some kind of standard, though we are not proposing a specific format here.

Lastly, traceable accounts of uncertainty allow identification of the Relevant Dominant Uncertainty (RDU), which may be of interest both to scientists (as a pointer towards fruitful avenues for further research) and to decision-makers (as an indication of where the uncertainty arises).

References and further reading

Mastrandrea, M. D., Field, C. B., Stocker, T. F., Edenhofer, O., Ebi, K. L., Frame, D. J., … & Plattner, G. K. (2010). Guidance note for lead authors of the IPCC fifth assessment report on consistent treatment of uncertainties.

Comments on Principle 5
Hailiang Du – P05-0726
Model discrepancy is often the major source of uncertainty, yet it is difficult to account for with reasonable reliability, let alone to trace them.

In my view, discrepancy is simply not well-defined in large-scale simulation modelling; the challenges include Lorenz’s “subtractability” and uniqueness; one model-state corresponds to an infinity of initial conditions and thus an infinity of outcomes outcomes. (LAS)

WHY?
For the consequences of uncertainty in scientific information to be understood and appropriately acted upon, the uncertainty itself must be communicated effectively. For numerate users in sectors such as insurance, this might include very quantitative measures of model variability. For users who are more concerned with preventive action, or the very risk-averse, an overall expert synthesis may be more appropriate. There is, of course, greater confidence in some outcomes and projections than in others.

TELL ME MORE
It is clear that a one-size-fits-all approach to the provision of uncertainty information is not appropriate; it will fail to satisfy the most numerate at the same time as it intimidates less numerate users into omitting any consideration of uncertainty at all. Rather than dropping to a lowest-common-denominator, or insisting that all consumers must learn statistical methods before taking into account any climate information, it makes sense to tailor uncertainty guidance for different contexts.

This is also a resource-constrained exercise: time spent on providing tailored uncertainty guidance is time not spent on conducting the original analysis; we contend, though, that uncertainty guidance is an integral part of an uncertainty analysis. It also requires definition of the audience, but if the consumers of the information are unknown then the utility of the whole exercise is questionable.

Comments on Principle 6
Dewi Le Bars – P06-0725
I don’t like “consumer of information”. This comes from economics consumerism. The information is not consumed, it is used. So I would prefer “user of information”.

There are objections to every possible word used here. Our use is not intended to have the negative connotations you note. Another alternative we often use is “practitioner.” Should a glossary ever come into being, we will make these distinctions more clear. Thank you for the comment, as it helps start that clarification. (LAS)

WHY?
Effective communication is best targeted using means of communication which have been demonstrated to be effective. There are long literatures in social science and psychology about the interpretation of numerical and descriptive statements and whether these match up with the intended interpretation. The best way to communicate the same insight to different target audiences is expected to differ, as is the best way to communicate different scientific insights to the same audience.

TELL ME MORE 
The IPCC have attempted to standardise communication about uncertainty ranges using a quantitative scale of likelihood and qualitative scale of confidence in that likelihood assessment (e.g. “likely (high confidence)”). The scale itself is not totally coherent, and its use in practice is patchy, inconsistent, and confusing to many readers. In addition, the statements to which the uncertainty language is attached need to be unambiguous in content and refer to a real-world variable (see Principle 4).

We would like to see greater use made of evidence from the literature in social science and psychology to inform design of more effective communication methods. Professional communicators (as opposed to professional scientists) may also have useful insights to share.

References and further reading

hard below
Budescu, D. V., Por, H. H., Broomell, S. B., & Smithson, M. (2014). The interpretation of IPCC probabilistic statements around the world. Nature Climate Change, 4(6), 508.

Bradley, R., Helgeson, C., & Hill, B. (2017). Climate change assessments: Confidence, probability, and decision. Philosophy of Science, 84(3), 500-522.

Comments on Principle 7
Luke Bevan – P07-2307
(also kind of applicable to 6 & 8) There is currently a decent literature about how people interpret risk-based or probabilistic information and how best to communicate this. There are also beneficial spillovers from other fields where risks must be communicated (e.g. medicine). What there is less of, is an understanding more generally about the communication of uncertain (non-risk) information. There have been some calls to professionalise/streamline the communication of climate information for lay audiences by using these risk-based techniques. However, this approach may run afoul of where people try to shoe-horn all uncertainties into being a form of ‘risk’ and leaving out deeper uncertainties.

WHY?

Where the other Principles are not followed, it is possible for the output of climate models to be interpreted with too much confidence. For example if we misidentify a model variable with a real variable, or where no traceable account is provided and some source of uncertainty is omitted. Having unwarranted confidence in climate information may lead to maladaptation, for example by investment in infrastructure which is optimised for a certain incorrect outcome. Misguiding decision-makers today will be exposed by future advancements of the science, if not by informed questions from them today.

TELL ME MORE
The “oversell” of climate information may be accidental, due to simple omission or lack of understanding of sources of uncertainty. It may alternatively be a result of poorly-aligned incentives; for example, we note that many research calls are aimed at “reducing uncertainty” where honest quantification may in fact result in increasing the stated range of uncertainty (due to previous underestimation). Funding is more likely to go to the more ambitious uncertainty-reducer, and consultancy jobs are more likely to go to the modeller who “can” give an answer at the requested 5km resolution than one who “can’t”.

References and further reading

Frigg, R., Smith, L.A. and Stainforth, D.A. 2015 ‘An assessment of the foundational assumptions in high-resolution climate projections: the case of UKCP09‘, Synthese. DOI: 10.1007/s11229-015-0739-8.

hard below
Frigg, R., Smith, L. A., & Stainforth, D. A. (2015) An assessment of the foundational assumptions in high-resolution climate projections: the case of UKCP09. Synthese, 192(12), 3979-4008.

Comments on Principle 8
Hailiang Du – P08-0726
Oversell science information is not common only in climate science but almost all science fields, which is damaging the whole academic creditability.

WHY?
Tuning or calibrating a model (making alterations to the parameters to improve the performance) should be done using only past observations and fundamental physical principles. Running the model forward then explores the space of what is possible given our modelling assumptions and consistency with past observations. If we constrain in addition based on future performance (for example, removing simulations which experience more than 0.2K/year of global mean temperature change, or those which fall outside our expectation of the climate sensitivity) then we artificially narrow the range of outcomes. To do so we must indeed be confident that the situation is unphysical, and understand how it can still be consistent with the model dynamics and our prior assumptions.

TELL ME MORE
“Future tuning” need not be a deliberate choice, but can occur accidentally. For example, it is common to re-tune models multiple times to improve the outputs. How do you know when to stop this process? If you do it once, run the simulation, and find that your previously well-behaved model now has a climate sensitivity well outside the range of the CMIP5 “pack”, do you re-tune? No doubt you find an error or a bug, and you re-tune it confidently, coming out with something more expected. But of course, the observation that your climate sensitivity is very high was the prompt, so this is future-tuning. It will always be possible to correct one more error or remove one more bug, but you are more likely to look for these when the model is not performing “as expected” and more likely to stop when it is doing what you expect. This could result in something elsewhere referred to as “intellectual phase-locking”, whereby the expectation becomes a (model-)reality and reinforces future expectations.

References and further reading

Hourdin, F., Mauritsen, T., Gettelman, A., Golaz, J. C., Balaji, V., Duan, Q., … & Rauser, F. (2017). The art and science of climate model tuning. Bulletin of the American Meteorological Society, 98(3), 589-602.

Schmidt, G. A., Bader, D., Donner, L. J., Elsaesser, G. S., Golaz, J. C., Hannay, C., … & Saha, S. (2017). Practice and philosophy of climate model tuning across six US modeling centers. Geoscientific model development, 10(9), 3207.

Mauritsen, T., Stevens, B., Roeckner, E., Crueger, T., Esch, M., Giorgetta, M., … & Mikolajewicz, U. (2012). Tuning the climate of a global model. Journal of advances in modeling Earth systems, 4(3).

Comments on Principle 9
Wilfran Moufouma-Okia – P09-0718
Not sure what is meant by “future-tuned simulations designed intentionally to illustrate some selected model property must be clearly distinguished from forecasts and projections (predictions) which, while conditioned on future forcing, are tuned using only the past” in the climate science realm? It is my understanding that for climate modellers once a model configuration/formulation is frozen/adopted in the development cycle, tuning will consist mainly in selecting parameter values in such a way that a measure of the discrepancy between observations and model outputs or between a modelled process and theory is minimized to a level acceptable in terms of process studies and process-oriented metrics. In this sense, tuning is often perceived as a way to compensate for model errors over the historic period for which observations exist.  

We agree that tuning using the past is good practise, and that further the traditional statistical good practice of not using the same historical data many times must be relaxed in climate-like simulations where one “cannot afford” to wait 50 years for truly out-of-sample observations. The objection here would be if one runs the model into the future, dislikes the result, and then tunes the model to achieve a model-future one finds more acceptable. In that case we question whether the model can be considered a forecast engine in any sense. (LAS)

WHY?

Good practice, when a system is stable and large datasets can be obtained, consists of robust out-of-sample evaluation of forecasts against real-world outcomes. In extrapolation tasks, this may simply be unachievable, either because few data are available for evaluation, or because the underlying system is changing in such a way that past performance may not be a reliable guide to future success. Good practice here includes acknowledgement of the limitations of statistical methods. Where “bad practice” is employed, for reasons which might include time and resource limitations and the incremental scientific value of conducting an incomplete analysis, it is important to identify what the impact may be on the results and signpost the nature of the incompleteness to any downstream customers.

TELL ME MORE
As an example, the IPCC’s headline projections of global mean temperature change in 2100, reported in their Summary for Policymakers, are based on climate models. Instead of reporting the 90% model range as a 90% (“very likely”) range for the real-world outcome, they report the 90% model range as a 66% (“likely”) range for the real-world outcome. This is good practice: acknowledging the limitations of models and giving a rough steer about the potential implications for decisions and further analysis based on these results. Unfortunately, it is rarely propagated into further analysis such as downscaling or impact modelling.

References and further reading

IPCC, 2013: Summary for Policymakers. In: Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change [Stocker, T.F., D. Qin, G.-K. Plattner, M. Tignor, S.K. Allen, J. Boschung, A. Nauels, Y. Xia, V. Bex and P.M. Midgley (eds.)]. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA.

Aims of the Lorentz Principles

How should the reliability of scientific findings be assessed and communicated to policy-makers? Different organizations, ranging from the IPCC to the CIA, have developed Uncertainty Guidances to assist experts in communicating uncertainties more clearly. Within many domains, however, such Uncertainty Guidance does not yet inform many numerate practitioners. This has restricted the development of a deeper understanding and characterization of what today’s science says, and hampered the communication of uncertainty in model-based findings to decision-makers in both the private and public sectors. The Lorentz Principles have been developed to meet that need, and the aim of this website is to facilitate further discussion, criticism, and refinement of the Lorentz Principles and their application.

Feeding back

We welcome constructive comments on the draft text of each Lorentz Principle and look forward to a discussion about the Principles, their application in practice, and their consequences.

Please connect with us to comment or chat via social media #LorentzPrinciples:
Leonard Smith: @lynyrdsmyth
Arthur Petersen: @ArthurCPetersen
Erica Thompson: @h4wkm0th

Where did the Lorentz Principles come from?

A Lorentz Center Workshop “Uncertainty Guidances in Science and Public Policy, Lorentz Center, Leiden, 13-17 November 2017” was convened in Leiden, the Netherlands, from 13-17 November 2017, with the aim to review existing Uncertainty Guidances and to propose ways forward for improved communication of uncertain climate information. The workshop brought together Uncertainty-Guidances-in-Science-and-Public-Policy-program 20 natural scientists, social scientists and philosophers – as well as practitioners who use scientific information to tackle real-world problems.The workshop developed draft principles for the responsible use, provision and design of scientific information on climate change for policy use and decision-making. Basically, the principles are to provide the background for discussing “good practice”, and bad, for questions of nontrivial extrapolation. These draft principles were subsequently refined and reviewed, resulting in ten Lorentz Principles for the Communication of Climate Information on 5 April 2019. We welcome your further input, discussion and refinement.

Where can I read more?

A list of relevant further reading is on our page of References. If you would like to suggest additions to this list, please contact us via social media.

The Lorentz Principles are supported by: 

Leonard Smith, LSE
Arthur Petersen, UCL
Erica Thompson, LSE
Luke Bevan, UCL
Seamus Bradley, University of Leeds
Mandeep Dhami, Middlesex University London
Nigel Harvey, UCL
Mike Hulme, University of Cambridge
Andrew Kruczkiewicz, IRI, Columbia University
Dewi le Bars, Royal Netherlands Meteorological Institute (KNMI)
Reason Machete, Botswana Institute for Technology Research and Innovation
Wilfran Moufouma-Okia, WMO
Magda Osman, QMUL