Minnesota Climate Explorer Frequently-Asked Questions

The Minnesota Climate Explorer has information icons and pop-up explanations to help users make decisions while using the tool. This FAQ is for those questions that cannot be addressed within the user interface. If you cannot find the information you seek, you may send a question to [email protected]e.mn.us.

 

What is the Minnesota Climate Explorer?

The Minnesota Climate Explorer is a simple, user-friendly climate information portal that allows users to graph and analyze past and projected future climate information at a variety of scales across Minnesota.

What is the difference between this tool and the Minnesota Climate Trends tool?

The Minnesota Climate Explorer can be thought of as a “new and improved” version of the Climate Trends tool that combines the functionality of the original tool with the ability inquire about projected future climate conditions for Minnesota. Everything in the “HISTORICAL” tab is, essentially, the Climate Trends tool, but with a newer platform that is compliant with current development standards. The “PROJECTED (FUTURE)” tab is new to the Climate Explorer and did not exist in the previous tool.

Why did you make these changes (to include climate projections)?

The Minnesota Climate Trends tool arose from a high volume of requests to have a locally-relevant climate data tool that would allow visualization and summarization of recent or ongoing climatic changes. The tool became very popular for assessment and planning purposes among resource managers, scientists, decision-makers, and data-curious users--many of whom expressed a desire to visualize Minnesota’s future climate as well. Therefore, in 2020, the Department of Natural Resources began redeveloping the portal to include future projections for Minnesota.

What will happen to the Climate Trends tool?

We want to give users a chance to get used to the new Climate Explorer before forcing them to give up the very popular Climate Trends tool. Therefore, the original Climate Trends portal will remain active into fall 2021 at least. We will notify users of the phase-out date at least 45 days in advance.

Why didn’t you keep the name “Climate Trends” in the new tool?

The term “trend” has a specific meaning in climate science, and indicates that some climate variable is changing over time in a manner that would pass statistical tests of “significance.” Minnesota’s climate is changing rapidly, but the changes are not equal across the state, across the seasons, and from one climate variable to the next. For instance, winter temperatures are increasing over time everywhere in the state, but average summertime high temperatures have been increasing in northern Minnesota, holding nearly steady in central parts of the state, and decreasing slightly in the south.

Some climate variables in some parts of the state and at some times of the year exhibit trends, but others exhibit no trends and instead only feature the usual ups and downs we expect from our extreme and variable climate. A given user may query historical data and find a trend, or not, depending on what is being sought. We have opted to keep our naming conventions focused on the data type, and not the interpretation of the results of a particular query.  Therefore the two portals inside the tool reflect the category of data: either historical, or projected (into the future).

What are climate projections? Are they the same as forecasts?

Climate projections come from climate models. We do not know exactly what the climate will be like in the future, but we understand its fundamental workings, and that increasingly, it is being influenced by excess greenhouse gases in the atmosphere (the air that surrounds us). Climate scientists have been able to measure and quantify many of the changes that have resulted.

Climate scientists use sophisticated mathematical models that can be run only on supercomputers to understand the likely range of future climatic conditions.  These "General Circulation Models", often referred to as "global climate models," combine the complex processes and relationships of our integrated ocean/earth/atmosphere system, with different scenarios for future greenhouse gas concentrations, along with other human and natural activity on the planet. It is only through these models that scientists can obtain research-based estimates of the ranges of possible conditions for the future.

The climate is complex and no single model can capture its behaviors perfectly. The atmosphere is three-dimensional and in constant motion, with continuously-changing conditions. Climatic models, therefore, are so complex and so computationally-demanding, that not all processes can be represented perfectly. The teams of scientists running each model have to decide which processes and properties will be represented most faithfully, which will be estimated more broadly, and how much "weight" the model will give to the various parameters and relationships it represents.

As part of a coordinated global research campaign, scientists at climate modeling centers and institutes around the world run different models with different process-weighting representations, and share their results through international, peer-reviewed scientific journals. Each unique model produces a massive array of modeled (or estimated) future climatic conditions around the world. More information is available at https://www.wcrp-climate.org/wgcm-cmip.   

Climate projections are not forecasts. The various estimates of future climatic conditions produced by models are known as projections, and should not be confused with forecasts or predictions. Climate models are designed to help us understand the general trajectories of climatic changes expected, based on greenhouse gas concentrations and other influences, but they are not designed to predict specific climatic conditions for a specific place and time. Thus, virtually all models agree that Minnesota will continue getting warmer in the future, but none can tell us with any certainty what exactly that means for the year 2046 (as one example).

What is the source of climate projections used in the Climate Explorer?

The climate projections used here were produced by University of Minnesota climate scientists, in support of two different projects: one on water resources, and another on agricultural pests.

These projections offer two basic benefits: 1) they were produced specifically for the state of Minnesota (whereas most climate projections cover larger regions, or even the US), and 2) they were localized for Minnesota by using a "downscaling" process that is based on physical relationships and atmospheric processes, rather than on statistical techniques. See the next question for more information about the differences between these techniques.

What is "downscaling," and why does the type of downscaling matter?

General Circulation Models (or global climate models) are global, with global-scale inputs and outputs. The purpose of these models always has been to help us understand how changes in greenhouse gas concentrations, along with other climate "forcings" and feedbacks, will affect our interwoven climate system. The outputs of these global models tend to have limited geographical detail, and tend to over-generalize or over-simplify local climates. In order to bring realistic climatic detail back to an area the size of Minnesota, the outputs from the global models must be "downscaled."

There are two basic approaches to downscaling: statistical, and dynamical. Statistical downscaling is computationally efficient, and can produce relatively "smooth," believable outputs, by fitting them to existing or theorized relationships between the model variables and the existing climate. The downside of this technique is that it is not physically based; the outputs are not responding to any atmospheric process but instead to a set of (potentially valid) statistical assumptions.

Dynamical downscaling, by contrast, re-introduces local atmospheric processes into the climate model outputs. This type of downscaling produces synthetic or modeled "weather," replete with pressure, wind, precipitating systems, clouds, cold fronts, warm fronts, and even individual thunderstorms.

Like global modeling, dynamical downscaling is computationally demanding and time-consuming, requiring extensive supercomputing resources. Scientists using these approaches also have to make decisions about which processes to emphasize most and which to generalize.  Scale is an important consideration also, because a very high-resolution project (high geographic detail) will require far more resources than one that is more "coarse."

The projections used in the Minnesota Climate Explorer were downscaled dynamically, at an "intermediate" resolution of approximately 10 km, which is fine enough to capture many of the processes that drive most precipitation events during the growing season, though too coarse to detect highly localized, or "microclimatic" influences . Even at this intermediate resolution, computational demands prevented the modelers from covering the entire 21st century, so they decided to obtain projections for distinct 20-year periods, which led to some differences in the graphing options between the Historical versus Projected time frames.

Why do the Historical and Projected graphs appear different?

The historical data portal (the Climate Trends tool functionality) is a climate time series-builder that produces one value per year for whichever time frame you select. For example, one July temperature value for each year, or one May – September precipitation total for each year. These can be plotted continuously for the range of years selected, because there is a value for every year.

The projected (future) values are discontinuous through the 21st century as a result of constraints of the project for which the modeling was done. So instead of having values for every year, we have values for two 20-year periods (one of which has two different greenhouse gas scenarios), plus one recent-historical 20-year period for comparison. Using continuous-time graphs is not an option, so for the projected values we chose simple vertical-line plots that represent the average and range of the 20 annual values per model for each 20-year period.

What do the Projected graphs mean? How do I interpret them?

These graphs display nine different projections per time period. These include one "Model Mean," representing the average of eight (8) individual dynamically-downscaled General Circulation Models, plus the outputs from each of the individual models.

Focus on the Model Mean plot, then use individual model plots to understand variations among the models. By far, the Model Mean, displayed in blue and on the left side of each time period, is the most important single plot to examine for any future period/scenario, because it contains information from all the other models. Individual models may run warmer, cooler, wetter, or drier than the others. These variations are normal consequences of the wide variety of modeling choices made by the different model teams. Users therefore should consider the mean to be the most reliable of the plots, and should consider differences from one model to the next to be examples of variations around the mean. Users should not fixate or depend on a single model projection plot for any time/period/scenario.

Each graph displays one past and two future time periods. Each graph covers one recent past period for comparison purposes (listed as "1980 - 1999 Modeled Present"), followed by a mid-century period (2040-2059), and two late-century (2080-2099) periods--one with lower and one with higher greenhouse gas concentrations. 

For each time period and model, the graph displays the average and "range" of values. Each time period has 20 years of data per model. To represent the key elements of those years efficiently, the plots display markers and vertical line segments. The marker within the line represents the mean (or average) value for that period. The vertical line segment simply represents the range of values, extending from the lowest to the highest. Thus, focus on the dots or markers, and use the line segments to understand the yearly variations each model (or the model mean) produced for a given time period/scenario.

 

 

*Using the "median" or mid-point is common practice for Box-and-Whisker plots, and is the preferred "central value" measure for climatologists working with periods shorter than 30 years. We use the mean (average) here, however, because those values were foundational to the "bias-corrections" that were applied to the downscaled model outputs. Our choice to use the mean instead of median only changes the vertical placement of the little block or dot, and does not alter the extent of the line segments on the graphs.

Why can’t I plot the drought indices in the projections portal? Why aren’t more climate variables available for the future?

Our first objective was to make the most basic projected climate variables for temperature and precipitation available, as these are by far the most frequently requested from users. This alone was a significant undertaking, with few examples to follow. Moreover, the data sources for the Historical and Projected portals are different. The “Palmer” drought variables available in the Historical portal were not used in the projects that sponsored the climate modeling and are not available to the Climate Explorer. Other projected climate variables that are available may be added to the Climate Explorer as time and resources permit.