Best practices for using climate projections
Topics covered in this article:
- What are some best practices for using climate projections?
- Understanding uncertainty
- About natural variability
- Guidance on using climate projections
What are some best practices for using climate projections?
Understanding uncertainty
Climate projections are our best approximations of future climate, but as with any statement about the future, there is no way to be certain they are accurate. One source of uncertainty in future climate projections is human greenhouse gas emissions. Projected climate data may not prove to be accurate if the actual emissions pathway we follow differs from the scenarios used to make the projections.
Another source of uncertainty in climate projections is the fact that different climate models – the tools used to simulate the climate system and produce future climate data – may produce different outcomes. There are more than 30 global climate models developed by climate modeling centers around the world, and they have different ways of representing aspects of the climate system. In addition, some aspects of the climate system are less well understood than others. Climate scientists are constantly working to improve our theories of the climate system and its representation in climate models. In the meantime, one way to account for model differences is to look at projections from as many different models as possible to get a range of possible outcomes. You can then take the average of the values across the different models, and this average value is a more likely outcome than the value from any single model. The default visualizations in Cal-Adapt are based on the average values from a variety of models.
It is important to note that here the term “uncertainty” is being used in the scientific sense, to acknowledge that there is a range in possible future outcomes. When we discuss sources of uncertainty in the global climate models, we do not mean we are unsure whether climate change will affect California. That climate change is occurring and is caused by human activity is the consensus of the overwhelming majority of scientists engaged with the issue. Learn more about the indicators of climate change in California. What is less certain is the extent to which the climate will change in the future, the pace of future climate change, and precisely how the changes will affect natural and human systems.
About natural variability
If you have lived in California for any length of time, you know that our current climate experiences a great deal of natural variability. Some summers are much hotter than others, some winters produce more snow than others, and some fire seasons produce more burning than others. Variability is a natural part of our climate systems and will not cease under climate change. Projections show that climate change will cause increases in average summer temperatures, but not every year looks like the average year.
In the future, as in the current climate, some years will be hotter and some will be cooler, but overall, future years will be warmer compared to our current climate. It is important to consider both long-term trends and extreme values (magnitude and frequency) in analyzing future climate variability.
![Chart, line chart, Natural variability in annual precipitation, Sacramento, 1960-2005](img/get-started/gs5_sac_hx_precip_lg.jpg)
Natural variability in annual precipitation, Sacramento, 1960-2005
Guidance on using climate projections
We recommend a few principles for working with climate projections:
1. Analyze data at a community scale, rather than looking at individual buildings, parcels, or grid cells.
Climate models have been proven to skillfully provide outputs that describe both past and future climate conditions. However, they are not capable of describing exact future climate conditions for small areas (like a single property parcel); they’re much better suited for community-scale analysis, even when model outputs have been downscaled.
2. Select longer time periods for more useful information.
Because future climate projections express natural climate variability, analyzing a longer time period gives you a better sense of overall future conditions. In other words, if you analyze just a few years of a future climate projection, you might happen to select years that are anomalous. You will get a more accurate picture of future conditions if you look at a period of at least a few decades.
Many scientific entities endorse using a thirty-year window for climate analysis so as to align with standard analyses of climatological normals. For example, within a thirty-year window, you can calculate average and extreme conditions using annual values as your dataset. You can also compute rolling thirty-year averages to study projected change through time.
Two thirty-year climate change adaptation planning periods are often used in California and are embedded in Cal-Adapt’s tools. These two periods align with those used in the Fourth National Climate Assessment.
- Mid-century: 2035 – 2064
- End-of-century: 2070 – 2099
3. Look at both model-averaged projections and individual models.
Climate models are most powerful for adaptation planning when they are used as an ensemble to describe the full range of plausible future climate conditions, since the average values across different model projections are considered more likely than any individual model value. However, you should also look at individual model values to capture the full range of model outcomes.
Cal-Adapt makes it easy for you to work with data from California’s priority models, as described in Section 3.1 of Climate, Drought, and Sea Level Rise Scenarios for California's Fourth Climate Change Assessment. The full suite of 32 downscaled GCMs is also available for download in NetCDF format.
4. Compare modeled projections to modeled historical data, rather than observed historical data.
The climate models that produced the data hosted on Cal-Adapt have all been validated, meaning that they have been confirmed to reconstruct past climate accurately. However, they don’t necessarily produce records of individual historical events, since climate models describe climate (long-term trends in weather), rather than weather (specific events).
To ensure that the climate change analysis you conduct is robust, climate scientists recommend that you compare modeled climate projections to modeled historical data – in other words, climate model hindcasts. That way, you’re working with like datasets to calculate climate anomalies.
5. Consider different greenhouse gas scenarios separately.
Since you know that averaging the results from different climate models is deemed more likely than any individual model, you might think this principle also applies to the different greenhouse gas scenarios. However, it is not the case that averaging together the greenhouse gas scenarios gives you a more likely greenhouse gas emissions pathway, and you should avoid doing this. Instead, think of each greenhouse gas scenario as a separate possible future.
Additional information on using climate change projections to inform adaptation planning can be found in California’s Adaptation Planning Guide, phase 2.