Close Window

Iowa State Univ.: "Summer forecast problems in the Chicago area"

Final Report


This Cooperative Project employed a high-resolution (8-km) mesoscale model over the domain of southern Lake Michigan to understand and operationally forecast the lake breeze. This phenomenon can directly affect over the 8 million people of northeast Illinois and far northwest Indiana both on land as well as over the water. A data feed was developed to input real time data from RAOBs to initialize the model. We also tested the use of Eta model output to drive the high-resolution model, but found that this approach was not generally useful because of the pronounced sensitivity of the lake breeze to small errors in the background flow.

New technologies including the WSR-88D, local mesonets and shore based SODAR (acoustical Doppler profiler) were used to monitor lake breeze progress, verify model output, study fine scale structure and better understand the mechanisms of this phenomenon.


2.1 Summary of overall project objective(s) and a description of project participants.

In the original proposal the research was cast mainly in terms of performing case studies for lake breeze events. With the acquisition of a suitable Unix workstation by the NWS investigators our focus shifted more to operational implementation of the model (though case studies also were simulated). This was a much more demanding exercise than the work originally proposed, and challenged all available resources including both personnel and equipment to their limits. However the much greater utility of an operational model as opposed to retrospective simulations clearly justified the additional effort.

While implications of lake breeze movement can be many, including air pollution and convective development, the primary emphasis was on using the model output to realistically anticipate wind shifts at O'Hare and Midway airports, both of which are the most heavily used in the world. Wind shifts associated with the lake breeze can have serious impact on airport operations due to runway adjustments.

Another objective of this project was to compile a rough climatology of lake breeze occurrence. Previous studies during the late 1960s and early 1970s made some effort in this direction but in our case we have documented two hourly (interval) specific data from various sites in northeast Illinois over a 4 season (April-Sept) span.

The PI at Iowa State was responsible for model programming, routing of data to the model and related software issues. He also advised NWS personnel regarding the mechanisms of lake breeze formation and structure, taking advantage of prior modeling and observational experience in this area.

The PI at the NWS was responsible for monitoring model performance, making it available to the forecast staff, detailing positive (and negative) environmental factors, collecting the above mentioned climate data (and doing some analysis) and doing the overall assessment of usefulness of this scheme.

2.2 Description of research/development accomplishments.

Prior to the beginning of this work the 'feel' for the daily occurrence of the lake breeze in northeast Illinois was poor. While it is often assumed that this event will occur, its relation to background flow, orientation around the coastline (can easily be skewed by background wind), better defined expectation with respect to the primary airports and to some extent overall structure was lacking.

In the process of doing work such as this the daily anticipation of the lake breeze was much better defined. The forecast staff was able to use model output to, if nothing else compare the simulation to the events that could be seen on the 88D or in surface observation. This alone heightened the sensitivity of how this phenomenon behaved under particular synoptic regimes.

For much of this project a series of inland towers were available across Lake County Illinois as well as a SODAR unit located on the Lake Michigan shore near the Zion Nuclear site. By using these data sources one could obtain a both horizontal and vertical sequence of lake breeze characteristics. The time/height output of the SODAR (profiler) was on a scale of 10s of meters in the vertical and 15 minutes (ave) over time. Thus a reasonable time-cross section up to 400-600 meters (AGL) allowed close examination of lake breeze structure. The 2, 5 and 15 nmi tower string across Lake County provided an excellent measuring stick for inland movement.

One of the model outputs was an E-W cross sectional plane through southern Lake Michigan. Even though this was simulation the stratification of the thermal field over the lake and its relationship to wind variation was enlightening. By the same token another output field revealed the convective mixing which occurs along the shoreline. Put in animation using NCAR Graphics both of these provided a good conceptual image of lake breeze related processes.

As well as the actual implementation of the code, substantial work was devoted to improving "user interface" aspects of the model. Automated scripts were set up so that the NWS staff can run the model for retrospective cases of interest. The ability of staff to run a high resolution mesoscale model on-site and at-will is unusual if not unique to this NWS office. We also constructed a Web type display for viewing the model output. Forecasters were able to view model output using familiar WWW browser software (e.g., Netscape or Internet Explorer) instead of more cumbersome plot programs.


Early in the project the University PI conducted a seminar at the NWS office regarding the lake breeze phenomena including driving mechanisms and characteristics. For the most part this has been a project, which has been of local interest. Our focus has been on the local northeast Illinois area, which includes a unique coastal configuration. Thus the project has not lent itself to inter-office (NWS) involvement or applicability.

The University PI attended a C-STAR workshop held October, 1997 in Silver Spring, MD where he described the project and related work to a larger audience including NWS administrators, forecasters and University researchers.


Arritt, R.W., D.P. Todey, C.J. Anderson and K.M. Labas, 1998: An operational high-resolution numerical model for the Lake Michigan lake breeze. Preprints, Second Conference on Coastal Atmospheric and Oceanic Predictions and Processes, American Meteorological Society, Phoenix, AZ, January 11-16, 1998.

Labas, K.M. and R.W. Arritt, 1998: Use of an operational mesoscale model to forecast and understand the lake breeze in northeast Illinois. 16th Conf. on Wea. Analysis and Fcstg., American Meteorological Society, Phoenix, AZ, January 1998, 132-134.

Laird, N.F., D.A.R. Kristovich, R.W. Arritt and K.M. Labas, 1998: Climatology of the lake and land breezes for the coastal region of Lake Michigan. Second Conference on Coastal Atmospheric and Oceanic Prediction and Processes, American Meteorological Society, Phoenix, AZ, January 1998.


5.1 University's perspective

One of the most significant benefits to the University was the "stress test" imposed by using the model on a routine day-by-day basis. This pointed out the strengths and weaknesses of the model, and led to several corrections and improvements. Another benefit was the motivation to produce better and clearer ways of presenting the model output. We would not have thought to create such output products as synthetic meteorograms or a Web interface without the needs of our NWS colleagues for better ways of interpreting the model results.

It should be noted that while most of our focus has been on working with the model, the project has also demonstrated the insights that can be gained from the use of sophisticated observations. We have found several cases where high-resolution visible satellite imagery from the NWS indicated a well-developed lake breeze even though surface wind observations were ambiguous.

Without doubt the greatest impediment to the project the slow and sometimes unreliable Internet connection between the University and the NWS office. As a result debugging or development sessions that should have taken an hour or less often consumed nearly an entire morning or afternoon.

5.2 NWS perspective

There is no question that this project has been a significant addition to the understanding of phenomena, which can have great impact on northeast Illinois. It has allowed a thorough and structured approach to a problem, which can be of daily concern to the forecast challenge here. The characteristics of the lake breeze in both orientation and movement have become clarified and more easily anticipated.

In using a model to simulate this very regionalized event we have obtained a better grasp of when numerical simulation is likely to succeed and, at least as important, fail. Like many numerical models knowing when they may be valid and when not can be as critical as the output.

As mentioned above we have also assembled climatology of the lake breeze though daily logs of assorted parameters. This data set can be and has been used by other researchers upon request.

One offshoot of this work has been the acquisition of a separate HP workstation upon which the model is run. In addition to this work we have been able to employ this machine for WATADS. It will also likely find use in AWIPS applications since it has the same general OS as AWIPS.