Authors:
Lace Padilla
According to a new report by the International Panel on Climate Change (IPCC) [1], every region on Earth is already affected by human-driven climate change. Due to the numerous and varied nature of climate change impacts, some regions, such as the Gulf Coast and Eastern Seaboard of the U.S., will experience more high-precipitation hurricanes, whereas others, such as the western U.S., will experience increased droughts and severe wildfires.
Squaring personal observations of the weather with long-term climate projections can be difficult. For example, a 1 degree Fahrenheit change in temperature is hard to detect in our daily lives. But we do notice when there is smoke in the air from wildfires or when the New York subway floods from a hurricane storm surge.
Understanding the risks associated with climate change where you live can be challenging because climate projections and extreme weather events have uncertainty. Researchers have found, however, that both the average person and experts trained in statistics commonly misunderstand even simple types of uncertainty [2].
Without understanding the uncertainty in climate projections, we may make poor long-term decisions about reducing our personal risk and fail to design systems to help the people most at risk. Understanding the uncertainty in climate projections will help us have appropriate confidence in how our actions today can curb the looming catastrophic effects of a changing climate.
In this column, we discuss how to understand climate change uncertainty with Roger Bales (https://faculty.ucmerced.edu/rbales/) and Andrew Kruczkiewicz (https://iri.columbia.edu/contact/staff-directory/andrew-kruczkiewicz/). Bales is a distinguished professor of engineering at UC Merced and an adjunct professor at UC Berkeley. He has conducted water- and climate-related research for more than 30 years, authored hundreds of peer-reviewed journal articles, and received millions of dollars in grant funding. He has led regional, national, and international measurement programs to understand climate change and contribute to climate solutions.
Andrew Kruczkiewicz is a senior researcher and faculty lecturer at Columbia University, where he is based at the International Research Institute for Climate and Society. He is also a science advisor for the Red Cross Red Crescent Climate Centre. Kruczkiewicz is an expert in communicating climate and weather forecasts to decision makers in the humanitarian sector.
Without understanding the uncertainty in climate projections, we may make poor long-term decisions.
Lace Padilla: Most people have experienced variability or uncertainty in the weather. How is uncertainty in climate the same, or different?
Roger Bales: Climate is the average of weather. There is a lot of variability in weather—day-today variability, seasonal variability, interannual variability, and spatial variability. Think of three timescales for predictions.
Depending on the region, weather forecasts typically have skill about five days out, as random factors enter into what weather occurs at a given place beyond that. Uncertainty in weather forecasts increases for each day further out in the forecast because of these random factors (Figure 1).
 |
Figure 1. Comparison of short-term weather forecasts in which there is more uncertainty when predicting the temperature seven days in the future versus one day in the future. The uncertainty in the forecast increases over time. |
Seasonal climate outlooks are the next scale up. These look one to 13 months ahead. For example, seasonal climate outlooks would ask whether the next three-month winter period has an equal probability of being in the middle (average), warmer, or cooler tercile of the past three full decades (Figure 2). For the U.S., one of the main uncertainties for this timescale is sea-surface temperature and pressure patterns (e.g., El Niño) and longer-term atmospheric circulation changes (e.g., the persistent high pressure off the West Coast of North America).
 |
Figure 2. Illustration of how seasonal climate outlooks can predict shifts in temperature distributions compared to the temperatures of the past three decades. |
The largest scale is multi-decadal climate projections, which is what is found in the recent IPCC report (Figure 3). Uncertainty in these arises for a couple of reasons:
 |
Figure 3. Visualization of a multi-decadal climate projection from the IPCC, showing a range of global temperature increase scenarios. |
- Range of scenarios: Projections depend on the emissions that will occur, so we have scenarios ranging from business as usual to elimination of greenhouse gas emissions.
- Model uncertainty: Equations in the models cannot represent all aspects of the factors affecting climate; therefore there is modeling uncertainty.
Note that since climate is the average of weather, when we project a shift in climate, there must also be a shift in weather. However, the models also project changes in the statistical distribution of weather events (e.g., fewer, more intense rainfall events without a change in total annual rainfall). There is uncertainty in both the statistical properties of weather in a warmer climate and the time series of interannual (e.g., how long and severe droughts will be) and subannual (e.g., magnitude and timing of storms and flooding) variability making up that changed climate.
LP: In addition to the different types of uncertainty associated with timescales, can climate models have more or less uncertainty across locations?
Andrew Kruczkiewicz: Yes. Areas have different modeling challenges due to missing data and a lack of quality data, among other issues. For example, satellite data allows for coverage across large geographic areas—sometimes global—but there isn't uniformity in quality across all areas of coverage. There are fairly significant regional deficits in the availability and quality of data for climate variables from place to place, with these "errors" propagating, and sometimes becoming exaggerated, in any datasets and/or products. This is one of the challenges of developing global datasets. As an example, if you were interested in rainfall forecasts or observations, the scale and value of a forecast will vary considerably from place to place because some locations will rely on satellite data while others use ground measurement networks.
A lot of this goes back to the concept of data availability. Unfortunately, the European and North American countries have more data. Therefore, there are better observations, better forecasts, and better projections. There is a lot that goes into data equity. I'm glad that in the past year or so we have made some progress and are talking about data equity, specifically with data availability and quality. But there is a lot more that needs to be done.
If we're talking about the root causes of uncertainty, many times we look at the processing of the data and at the potential biases and errors within the computational steps. There is uncertainty that will come from that. But also, if we go back and see why some places have data and some don't, a lot of it goes back to environmental justice issues, too, and certain areas being ignored because someone thought they were less important.
Unfortunately, many of the areas perceived as less important are where the most vulnerable populations live. I think that's an interesting angle to look at uncertainty, too. What are the drivers of uncertainty where vulnerable populations live? Are there opportunities to address specific uncertainties? Where should we prioritize making sure uncertainty is not only lowered but also communicated so vulnerable populations are aware of what it means?
LP: What do you wish that people understood about climate uncertainty?
AK: I am trying to think more about how to normalize discussions around uncertainty and turn it from a bad word to an approachable word that we must live with. I think about this from some of the discussions around adaptation to climate change; we have to begin to live with disasters, live with floods, live with drier periods—and also live alongside scenarios where those events combine and occur in close succession (to learn more about uncertainty in compound hazards, see [3]).
Uncertainty leads people to be uncomfortable to some degree, or maybe feel like they aren't smart enough to process it or address it. It also introduces the concept of mystery or feeling out of control.
From my experience with the Red Cross, I was privileged to talk with farmers and herdspeople in Mongolia, an experience that made me think a lot about this. Mongolia is a place that has been dealing with extreme events for a very long time. They know that extreme weather is inherently uncertain, and within their lifestyle, they have built it into their systems. I guess they wouldn't call it understanding uncertainty or living with uncertainty, but it's something that they are addressing in more productive ways to think about how to embrace uncertainty, to inform what actions are taken. It's something like embracing uncertainty; it's not a bad word.
Unfortunately, many of the areas perceived as less important are where the most vulnerable populations live.
If we don't understand uncertainty, if we don't acknowledge uncertainty, then we won't be able to understand risk, and thus we won't be able to really do much when it comes to addressing climate change.
The IPCC's report is an excellent resource for learning more about climate change in your region. The IPCC leads the way in operationalizing and clearly defining the use of uncertainty terms such as extremely likely and high confidence (see Uncertainties Guidance Note from the fifth report [https://www.ipcc.ch/site/assets/uploads/2017/08/AR5_Uncertainty_Guidance_Note.pdf], which is under development for the sixth report).
The IPCC also released an interactive online tool, the IPCC WGI Interactive Atlas (https://interactive-atlas.ipcc.ch/; see Figure 4), that allows you to explore the climate data from the sixth report. With it, you can learn more about global trends and climate projections with uncertainty in your region.
 |
Figure 4. Screenshot of the IPCC's WGI Interactive Atlas, which allows users to explore IPCC's climate projections for regions around the world. |
References
1. IPCC, 2021. Climate Change2021: The Physical Science Basis. Contribution of Working Group I to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change. V. Masson-Delmotte et al., eds. Cambridge Univ. Press, 2021.
2. Belia, S., Fidler, F., Williams, J., and Cumming, G. Researchers misunderstand confidence intervals and standard error bars. Psychological Methods 10, 4 (2005), 389–396.
3. Padilla, L., Dryhurst, S., Hosseinpour, H., and Kruczkiewicz, A. Multiple hazard uncertainty visualization challenges and paths forward. Frontiers in Psychology 12, (2021).
Author
Lace Padilla is an assistant professor at UC Merced who studies decision making with uncertainty visualizations. Her work focuses on improving forecast visualizations to help people make high-risk decisions, such as hurricane evacuation or managing Covid-19 risks. She is also an advocate for diversity, serving on the IEEE VIS diversity committee and the Spark Society governing board. [email protected]
Copyright held by author
The Digital Library is published by the Association for Computing Machinery. Copyright © 2022 ACM, Inc.
Post Comment
No Comments Found