Exploring Uncertainties

Users are encouraged to explore the sensitivity of the Time of Emergence results to different assumptions and scenarios.  The selection options currently available in this tool are summarized below:

Table 1. User Selection Options
 Input Parameter Options
Emission Scenario
  • Higher (RCP8.5 or SRES A1B)
  • Lower (RCP4.5 or SRES B1)
Management Sensitivity
  • High (impacts triggered by extreme 20% high or 20% low of 1950-1999 conditions)
  • Low (impacts triggered by extreme 5% high or 5% low of 1950-1999 conditions)
Rate of Climate Change
  • Fast
  • Moderate
  • Slow

These options, together with the use of multiple global climate models, allow users to explore two main types of uncertainty associated with the Time of Emergence results:

  • Climate modeling uncertainties
  • Uncertainties arising from the estimation of Time of Emergence

The table below provides some guidance on which results one may wish to put more emphasis on, depending on the risk tolerance of the user or management context.

  Lower Risk Tolerance Higher Risk Tolerance
Emission Scenario Higher (RCP8.5 or SRES A1B) Lower (RCP4.5 or SRES B1)
GCM Agreement Low (25%) High (75%)
Management Sensitivity

Low (impacts triggered by extreme 5% high or 5% low of 1950-1999 conditions)

High (impacts triggered by extreme 20% high or 20% low of 1950-1999 conditions)

Rate of Climate Change Fast Slow

Lower risk tolerance: This may include users or management contexts that are risk averse or would choose to operate more conservatively in order to minimize the risk of potential negative impacts. An example might be when planning large infrastructure investments with long asset life which would be extremely difficult to relocate or retrofit if future conditions were worse than expected.  Selections in this category tend to yield earlier Time of Emergence.

Higher risk tolerance: This may include users or management contexts that are more flexible or could plan response actions that are relatively easy to implement). Selections in this category tend to yield later Time of Emergence.

Learn more about exploring uncertainty with this tool

Climate Modeling Uncertainties

Users can explore the effects of climate modeling uncertainties on the Time of Emergence results through the application of different:

  • Emission scenarios
  • Global climate models
  • Downscaled climate datasets

Emission Scenarios

Multiple emission scenarios are often used to represent forcing uncertainty, which captures things in the future that are considered outside the climate system per se, yet affect it (e.g., future population and land-use changes). Results from a single projection only shows a single plausible future, which overlooks the full range of possible outcomes.  Therefore, a multi-scenario approach should be adopted.

Because the underlying climate datasets used in this tool are derived from two sets of global climate modeling efforts through the Coupled Model Intercomparison Project (CMIP5 and CMIP3), the Time of Emergence projections offered here are based on two sets of emission scenarios: IPCC SRES (Special Report on Emissions Scenarios) scenarios and RCPs (Representative Concentration Pathways), respectively.

The Time of Emergence results are available for one high, and one low, emission scenario from each of these sets as described below. The “High” scenario is based on rapid greenhouse gas emissions with little to no mitigation strategies and “business as usual” approach to energy usage, which implies an earlier Time of Emergence estimate due to greater effects of climate change; “Low” is based on lower emissions, a high level of mitigation strategies, and use of alternative energies, and implies a later Time of Emergence estimate.

Table 2. Emission scenarios available in this tool.
  IPCC SRES (CMIP3 results) RCPs (CMIP5 results)

These reflect changes in the way economies are structured, population grows, technology develops, as well as energy intensity and land use changes. None of them includes a scenario where action is taken to drastically reduce GHG emissions (known as a stabilization scenario). They are presented as four equally plausible (note: this does not mean equally likely, as SRES did not assign probabilities) storylines labeled A1, A2, B1 and B2. They represent different world futures in two dimensions: a focus on economic (the “A” scenarios) or environmental (the “B” scenarios) concerns, and global or regional development patterns.

RCPs represent the full range of potential future radiative forcing pathways that are considered to be feasible and are compatible with the full range of stabilization, mitigation and baseline emission scenarios available in the scientific literature.

Development Approach Developed sequentially, i.e., from detailed socio-economic storylines that determine GHG emissions to radiative forcing. Developed through the parallel approach, where important characteristics for scenarios of radiative forcings, such as the level of radiative forcing in the year 2100, was first identified.
Scenarios used in this tool
  • High SRES scenario: based on SRES A1B, labeled as “SRES High”
  • Low SRES scenario: based on SRES B1, labeled as “SRES Low” 
  • High RCP: based on RCP8.5, labeled as “RCP High”
  • Low RCP: based on RCP4.5, labeled as “RCP Low”
References Nakićenović et al., 2000 Moss et al., 2010; Vuuren et al, 20

Global Climate Models

Multiple global climate models are often used to represent structural uncertainty. Even the most sophisticated models are unrealistic representations of many relevant aspects of the climate system due to our limited understanding of, and inability to simulate, the Earth’s climate, such as grid resolution and missing, omitted or approximated processes that cannot be accurately described in the model. What processes to be included in a model and how they are represented may be subjectively chosen based on expert knowledge and experience by different modeling groups; this constitute structural uncertainty.  Since there is no single “best” global climate model, a multi-model approach is recommended.

This tool presents Time of Emergence results derived from projections of 21 and 6-7 global climate models from the CMIP5 and CMIP3 experiments, respectively (temperature, precipitation and hydrologic projections were available from six CMIP3 models; streamflow projections were available from seven CMIP3 models). Although selecting a particular, or a subset of, global climate models is not an option, users can explore the central tendency, spread and level of agreement in the results due to multiple global climate models through the different features available in this tool:

Tool Features Representation of Multi-Model Results
EXPLORE by Locale  
Timeline Indicates multi-model median Time of Emergence
Summary Table Reports central 50th percentile Time of Emergence range
Boxplots Show (graphically) multi-model median Time of Emergence, 25-75th percentile Time of Emergence range, and individual GCM results
EXPLORE by Variable  
Maps of Emergence Year Show multi-model median Time of Emergence
Maps of Emergence Location (by year) Show locations with 25, 50, 75% model agreement that emergence has occurred

For variables where different global climate models indicate different directions for the climate change signal, the signal direction is identified as the direction projected by 60% or more of the models. Time of Emergence results are computed and reported using only that subset of (60% or more) of models.


Downscaled Climate Datasets

Global climate models simulate future climate for the globe at a resolution too coarse to provide useful information for regional- or local-scale decision-making. Downscaling increases the spatial and/or temporal resolution of global-scale climate projections to provide detail relevant to regional/local management or operations. Different downscaling methodologies have varying strengths and weaknesses; the skill of different approaches varies temporally and spatially, and with variables. No single approach is therefore more superior, hence results from multiple methodologies should be used to explore downscaling uncertainty.

Note: Precision does not necessarily imply accuracy. Finer modeling resolution does not necessarily imply greater confidence or certainty, particularly regarding extreme events. For example, regional climate models still cannot capture all the important physical processes responsible for precipitation despite their finer resolution. Regional simulations also inherit the limitations associated with the parent global climate model. Thus, the process of downscaling adds another source of uncertainty to the projections.

Results based on two main types of downscaling approaches are available in this tool, including those from dynamical downscaling and statistical downscaling (see Input Datasets). Users are able to filter Time of Emergence results by downscaling method for the maps of Emergence Year and Emergence Location.


Time of Emergence Estimation Uncertainties

Users can explore the uncertainties arising from the methodology used for determining Time of Emergence through the definitions of the emergence threshold and error in estimating the climate change “signal” (see What is Time of Emergence?).

Definitions of the Emergence Threshold

The threshold at which “emergence” occurs reflects societal or natural system sensitivity to changes in that variable. This sensitivity will vary for different users and systems depending on their capacity to adapt to or accommodate change.

This uncertainty is incorporated in the Time of Emergence calculation through two user-selected levels of management sensitivity to change. The Time of Emergence results are available for one high and one low level of management sensitivity to changes in conditions:

  • High management sensitivity: impacts triggered by the extreme 20% high or 20% low conditions experienced during the baseline period (1950-1999)
  • Low management sensitivity: impacts triggered by the extreme 5% high or 5% low, conditions during 1950-1999


Error in Estimating the Climate Change Signal

This is the uncertainty in statistically estimating the climate change signal, i.e., the slope of the linear fit to the simulated data for the 21st century. The projected future time series of a variable (e.g., temperature) typically includes a steady trend, and fluctuations around that trend known as variability. The steady trend is assumed to be the climate system response to external greenhouse gas forcing and the fluctuations result from the various modes of internal climate variability. In computing the trend, one can place the true slope within statistical confidence limits depending on the strength of the trend compared to the variance in the data using a Student’s t test. Thus, Time of Emergence can be computed from each model using a high, central, or low value for the signal based on the confidence interval for the computed trend.

This tool provides Time of Emergence results based on three estimates for the rate of climate change (i.e., the slope of the simulated climate change signal) – the central estimate, along with the “low” and “high” estimates that define the 90% confidence range. Therefore, there is a 5% chance the slope is above the faster rate and a 5% chance it is below the slower rate.

In EXPLORE by locale, users can examine Time of Emergence results by estimated high/moderate/low rate of climate change. In EXPLORE by variable, results for the moderate rate of climate change are provided.


Continue Reading:



The Time of Emergence project was conceived and funded by U.S. Army Corps of Engineers Climate Preparedness & Resilience programs & U.S. Environmental Protection Agency-Region 10. Methodologies and stakeholder engagement were developed and implemented by the University of Washington's Climate Impacts Group. The Time of Emergence online tool was developed with support from the Center for Data Science, University of Washington-Tacoma.