Modeling Polar Climate

Real-time and Forecast Weather for the Arctic at the Polar Frontier Exhibit

When the concept of an exhibit like the Polar Frontier was first being discussed for the Columbus Zoo and Aquarium, the Byrd Polar Research Center was invited to help the public understand how the Arctic is responding to climate change. From the outset, the planning committee thought it would be great if visitors to the Zoo could see the kinds of conditions that the different forms of wildlife endure throughout the year in the “high latitudes”, compared to what we experience in Columbus.

Left Photo: Taken 08/21/2009 Patrick Kelley, U.S. Coast Guard, USGS Right Photo: Taken from The Ohio State University website

The Polar Meteorology Group at BPRC had already developed a version of a weather forecast model that is unique to the Polar Regions. This model is used daily by the National Weather Service and the U.S. Antarctic Program (part of the National Science Foundation). With additional support from the Ohio State Office of Outreach and Engagement and matching funds from Battelle Memorial Institute (Battelle), a real-time display of Arctic conditions with a 48-hour forecast was developed at BPRC to inform the Zoo’s guests. The model outputs are archived on the BPRC server so that they can be made accessible to anyone who is interested in following the weather systems as they track across the Arctic, whether as a hobby or for a research project.

Modeling Polar Climate at BPRC

Based on previous research and experience with mesoscale modeling, the Polar Meteorology Group (PMG) of the Byrd Polar Research Center (BPRC) at The Ohio State University has developed a polar version of the Weather Research and Forecasting Model (WRF). The National Weather Service uses the WRF model for daily forecasts in the continental U.S., but the Polar Regions have fundamentally different conditions that must be considered. For example, the type and extent of vegetation, the temperature of the land surface (and sea surface), and the distinct differences in elevation and characteristics of the North and South Polar Regions require a different range of possibilities than are needed to develop forecasts for the “lower 48 states”.

Polar WRF (pronounced like "wharf") is a high-resolution model designed specifically for use in Polar Regions. Short term weather forecasts generated by Polar WRF have been evaluated over Greenland, the Arctic Ocean, and Arctic land surfaces. Polar WRF is also used by forecasters as part of the National Science Foundation-sponsored Antarctic Mesoscale Prediction System (AMPS) to meet the logistical needs of the United States Antarctic Program (USAP). For this purpose, the model generates forecasts for the U.S. Antarctic Program to enable real-time critical decisions, for example to allow or cancel flights from New Zealand to McMurdo Station or from McMurdo to South Pole Station.

Flight path from Christchurch, New Zealand, to McMurdo Station, Antarctica

In addition, researchers who study the Arctic have recently decided that it is important to have a combined assessment of what is known about the Arctic region. The resulting effort, known as the “Arctic System Reanalysis (ASR) Project”, combines historical atmospheric, oceanic, land surface, and ice observations with field campaign observations and regional model output to produce a more complete description of the Arctic system, including the interactions of the atmosphere, sea ice, and land. The resulting description of the atmosphere-sea ice-land Arctic system will better allow researchers to detect changes in the coupled Arctic climate system. As part of this integration effort, Polar WRF will be used as the primary mesoscale model for the ASR.

What is the model structure?

In modeling, the area of focus is broken down into a grid consisting of points in both the horizontal and vertical directions. So a grid, like a checkerboard with defined square areas, also has a third dimension (height) to accommodate changes in the air that occur at different altitudes above Earth’s surface. Mathematical computations are performed within each 3-dimensional grid cube, in order to predict weather conditions within that specific cube. As a result, each cube in the grid system has a single value, and that value represents the conditions for the entire cube. Any point in the cube’s space would be given the same numerical value.

Graphic is an example of the model structure of a Global Atmospheric Model. Taken from NOAA.gov

What is "Model Resolution"?

Even though weather can vary greatly over relatively short distances, the model is constrained by the distance between grid points. As mentioned above, conditions throughout the entire grid cube are the same. Resolution refers to the number of grid points—the amount of data given. More grid points within an area of focus yields higher model resolution and in theory, increased accuracy of predictions. Lower resolution models have fewer data points per unit of area. This means that the grid points are farther apart or that there are fewer data points available.

(maps by Gary Strand, UCAR; used with permission)

What does "mesoscale" mean?

Mesoscale meteorology refers to the study of weather systems that range in size from a few kilometers to hundreds of kilometers wide. Mesoscale models, like the Polar WRF, are developed to have higher resolution within a limited area of focus. This means that more data points are included in the area of focus. For the Polar Frontier exhibit, the area of focus is the Arctic region. Keep in mind that larger weather systems can also be studied on a synoptic scale, ranging from hundreds to thousands of kilometers in size; but the model of the Arctic Region designed for the Polar Frontier is still an example of a mesoscale model.

Polar WRF Output for the Zoo's Polar Frontier Exhibit

As part of the research and development of Polar WRF for ASR purposes, a coarse (lower-resolution) version of Polar WRF is used to study the broad scale behavior of the atmosphere in the model, throughout the entire Arctic and surrounding latitudes. Included in the surrounding latitudes are important river drainage basins which are fundamental to the Arctic hydrologic cycle. This version of Polar WRF creates short-term (96-hour) forecasts of the Arctic region for a graphic display at the Polar Frontier exhibit.

The domain (or area) of the Polar WRF model is centered near the North Pole, with the boundaries located just inside the highest terrain of the Tibetan Plateau. It also includes a large portion of the North Pacific and North Atlantic storm track areas. The domain extends over the Great Lakes region and also includes all of Japan. The goal for this display is to demonstrate that the weather and climate of the areas surrounding the Arctic region are very much connected to the changing Arctic climate system. It was designed to encourage people of all ages to watch what happens in the Arctic and to become familiar with (and to test ideas about) how the atmosphere responds to day-to-day and seasonal changes.

Image created by Aaron B. Wilson, BPRC Polar Meteorology Group

Polar WRF has several integral parts in its configuration. First, the land surface must be specified throughout all areas of the domain. This is necessary to enable accurate and realistic interactions between the atmosphere and Earth’s surface within the model. Next, the model must be given a set of initial conditions and conditions along its boundaries. The boundaries include the land surface and soil conditions, sea surface temperatures and sea ice conditions, and atmospheric variables along the edge of the area of focus (the outer limits of the domain). For this purpose, information such as temperature and surface pressure is supplied from a global forecast model. In this case, the information is obtained from the National Centers for Environmental Prediction (NCEP) Global Forecast System (GFS).In addition, instrumental data, including satellite measurements of energy obtained from passive microwave radiometers, are input into formulas for known mathematical relationships that can be used to calculate the sea ice concentration. With all this information integrated, powerful computers are used to run the model and generate forecasts for the Arctic (the domain of the model).

The output shows the Arctic in a “now” version, or the “real-time” initial conditions. Then it generates a forecast representing the model’s best predictions for the region for up to 96 hours into the future. Students (or anyone interested in the current or predicted conditions for the Arctic) can check how the conditions have changed every 12 hours. They can also assess how well the model is doing in predicting the changes. The model output generated for the Polar Frontier will also allow those who are interested in watching intense weather systems as they move toward Europe or travel southward across Canada toward the Midwest. One type of weather system that may be tracked using Polar WRF forecasts are “Alberta Clippers”, small systems that form in Alberta Canada and advance into the Great Lakes region in the winter. They are usually responsible for light dustings of snow in Ohio, and cold air associated with high pressure behind the system is often reinforced following the storm.

The model developed for the Polar Frontier was brought online in May, 2010. Archived images from the model since that time are available and will be saved for research purposes. There are two types of archived images. The first provides a basic weather map including temperature and sea surface temperature. The second reflects the low and high pressure systems throughout the domain. Combined, these images will provide students the opportunity to observe changes in the Arctic as the seasons change and the archive documents the onset of the polar winter.