Example Applications

Example Applications of Analytics Engine Notebooks


The Cal-Adapt: Analytics Engine (AE) is a web based climate data platform that analyzes historical and future conditions in context using the next generation of climate data. Through this work, the AE increases capacity for analysts and decision makers, especially in the energy sector, in preparing for extreme weather events, ongoing climate change, and understanding future warming conditions. The following are some methodologies to utilize the existing notebooks in practical application:

  • Data Localization using the Getting Started Notebook
  • Extreme Weather Events using the Threshold Exceedance Notebook
  • Warming Levels using the Warming Levels Notebook

Data Localization

Understanding the changes to local climate supports decision making regarding asset management and ratepayer needs.

The Analytics Engine provides data localization capabilities to study and understand weather patterns at a local scale, by region, county, or city. Current energy forecasting methodologies rely on generalized data for a larger geographic area. With a dataset that covers a large geographic area, forecasting weather conditions for a specific location is difficult. The Python functions developed as part of AE provide paths to overcome that difficulty and understand specific projections.

For example, decision makers at a utility company traditionally use an existing model that produces electricity load forecasts based on historical weather observations. With Analytics Engine outputs, decision makers can now determine what upcoming climate conditions will affect load forecasting in a specific region. Functions within the Analytics Engine’s Getting Started notebook demonstrate how to bias-correct gridded model observations at a weather station using the weather station’s historical record and can guide users in:

  • Identifying a time period and emission scenario
  • Selecting weather station(s) in the location of interest
  • Localizing gridded data to a weather station(s)
  • Exporting for use outside of the Analytics Engine in an existing workflow

By using the time period, emissions scenario, and weather stations of interest AE can project air temperature and bias-correct the gridded model observations at that weather station. This notebook is fully customizable and offers visualizations of the data so that users can observe the overall pattern of and trends in bias-corrected air temperature within the AE models.

To learn more about the Analytics Engine’s specific localization methodology, the localization methodology notebook walks through the quantile delta mapping (QDM) process of how the localized data is developed for a weather station . This notebook provides step by step instructions on the localization process within the Select tool of the Getting Started notebook.

Extreme Weather Events

Exploring the extremes in temperature allows for the prioritization of temperature mitigation resources for both extreme heat and extreme cold.

The Analytics Engine provides the Threshold Exceedance notebook to identify the number of events that would surpass a threshold (e.g. 115° F) to proactively make decisions regarding asset management, including mitigation, maintenance, and design upgrades. For example, with the Threshold Exceedance notebook, an engineer could reliably estimate the number of extreme heat events in air temperature for a specific location to better prioritize equipment design and future siting of asset locations.

Within the Threshold Exceedance notebook, a user can:

  • Select a geographic area and time period of interest
  • Select a variable of interest (e.g., air temperature, precipitation)
  • Define the extreme event (threshold, recurrence interval, and duration of event*). The event exceedance plot thus illustrates the projected frequency of extreme events at the selected threshold and duration and can also be used in additional models or analysis.

Warming Levels

An approach to separate timing from the impacts to climate and make dynamic plans based on the amount of warming in the world.

The Analytics Engine’s Warming Levels notebook provides detail on the regional response to a given level of global warming, and illustrates how to overlay user data to help identify vulnerable locations. As an example, the Warming Levels notebook looks at power plants in connection to the regional response at a specific warming level. With the Warming Levels notebook, users are able to:

  • Visualize and compare the differences in the regional response across models at four different warming levels (1.5˚, 2˚, 3˚, and 4˚C of warming)
  • Export data at a warming level across models for specific application needs

In the Warming Levels notebook, a user would first select a geographic region and warming level to analyze the range of responses across the different models. Users would then be able to visualize the change in the selected climate variable from the historical climate at the designated warming level, with options to see different warming levels. In other words, it would show how much warmer, with temperature selected, a specific county would be when the world is 2˚ warmer.

With this information, a city planner could identify the local impacts of a warming trajectory and build plans to support the needs of that community at that warming level. Understanding warming levels at the regional scale can help inform regional policy decisions. For example, if more precipitation occurs at a certain warming level, policies around stormwater or flooding response can be planned and implemented in preparation of reaching that warming level.

It’s very important to be able to build plans for the scenarios (i.e., a specific warming level) and not the time (i.e., 50 years from now) because the timing of various climate impacts is highly variable due to model uncertainty and human actions, and ultimately focus on preparing for a likely scenario regardless of timing.