Close Window

Iowa State Univ.: "Utility of ACARS data in forecasting the Chicago Lake breeze"

Final Report


This project was an outgrowth of a COMET Cooperative project that focused on developing a mesoscale model to help predict the evolution of the lake breeze off Lake Michigan in real time. The intent was both as an aid to forecasting as well as helping to understand lake breeze mechanisms and climatology. The forecast problems for WFO LOT, which would be impacted, include a major metropolitan area (almost 10 million people) as well as two major airports. The latter are critically dependent upon wind direction for operational runways and require as detailed forecasting as possible for smooth transitions due to lake breeze passages.

Initial data sources for this model were NWS upper air sites located at Quad Cities (DVN), Green Bay Wisc (GRB) and Lincoln Illinois (ILX). These sites are all at least 200 km away from the area of interest and have data available only for the standard 00 UTC and 12 UTC reporting times. While model output appeared to work reasonably well given these restrictions, sometimes the forecasts were unrealistic because the lake breeze is very sensitive to the speed and direction of the background wind. The present follow-up study was intended to improve forecast reliability by using aircraft ascent/descent (ACARS) data for model input. Preliminary studies from the previous season suggested that model input providing more representative initial conditions (in terms spatial and temporal proximity to the target region) could provide more accurate lake breeze forecasts in 10-20% of the applicable days. Accordingly, we designed the Partners project to study the feasibility of using ACARS input from O'Hare International Airport (ORD) and Rockford, Illinois (RFD; a large center for UPS cargo flights).

Technical problems needed to be overcome to access the ACARS data and ingest it into the model. This consumed a great deal of initial resources and time. The generation of the model was also shifted from the NWS to Iowa State to better facilitate ACARS data ingest and allow running of the model on a LINUX platform. Prior to this output generation was done on a HP-J200 workstation using UNIX at the WFO. Output is available via the Internet. Given the experimental nature of the methodology, access to model output is restricted to sites in the NWS Central Region.

In the course of doing this work a few model adjustments were made to enhance the quality of the results. One of these was the need to smooth out boundary layer wind profiles to compensate for the nocturnal accelerations that occur due to surface inversions. This model seems to be quite sensitive to winds around 1000 to 2000 ft AGL. Increases in nocturnal wind speeds near the top of the surface inversion translate into an unrealistically strong geostrophic wind that alters the characteristics of the ambient background flow upon which the lake breeze later interacts. Smoothing of the wind profile reflects the mixing that would occur with daytime heating and better reflects the real environment later in the day.

Collaboration between the ISU and NWS investigators will not end with the official termination of the project. Since the procedure for running the model is now fully automated, the ISU group will continue to run the model in 2001, beyond the end of the Partners project. The model will be run indefinitely, as long as the data stream remains available.


While much of this work was facilitated via the phone, email and Web, Labas made one trip to Iowa Sate in the autumn of 1999 and Arritt made two trips to the WFO. One trip to the WFO was in part necessitated by the failure of one of the hard drives on the HP-J200 which resulted in the loss of the ability to locally run the lake breeze model on demand. The model as well as NCAR Graphics needed to be reinstalled.


Laird, N.F., D.A.R. Kristovich, X.-Z. Liang, R.W. Arritt and K. Labas, 2000: Lake Michigan lake breezes: Climatology, local forcing, and synoptic environment. J. Appl. Meteor. (in press).

Labas, K., and R.W. Arritt, 1999: Use of ACARS soundings to improve lake breeze forecasts. Third Conference on Coastal Atmospheric and Oceanic Prediction and Processes, American Meteorological Society, New Orleans, November 1999.


University Perspective

Our access to ACARS data through this collaboration has offered new insights. One is an appreciation for how much the atmosphere can vary over a rather short change in space or time, and the effect that this has on predictions from the mesoscale model. It is especially interesting to evaluate results from our sensitivity experiments regarding the effect of large-scale conditions on lake breezes as compared with day-to-day observations.

Problems mentioned in Section 1 were overcome and we now have developed a fully automated procedure for running the model both from ACARS data and from conventional sounding data.

Material from this project has been adopted in courses taught by the PI. For example, in AGRON 406 (a survey of world climates), our unit on microclimates now uses lake breeze imagery from the Chicago NWS radar to illustrate the importance of proximity to water bodies in controlling local climate.

Forecast Office Perspective

The undertaking of this project, in both the current Partners as well as the Cooperative version, focused the station's attention on the lake breeze. Daily warm season logs were maintained for 4 years which were used to verify model performance and have been subsequently used by at least on other researcher for Lake Michigan lake breeze studies. By maintaining a daily sensitivity to both model output and resulting lake breeze development we were able to obtain a more confident feel for how this phenomena would evolve each day (that it occurred). Differing background flows affected the lake breeze front is various was and this became a 'predictable' trait. These logs were also used to develop a rough climatology (frequency) for the lake breeze over the restricted area of northeast Illinois.

In addition we were able to employ the WSR-88D to monitor lake breeze progress as well as access (through 1998) a mesoscale network and SODAR unit along the lake shore over Lake County Illinois, just south of the Wisconsin border. The SODAR unit detected winds to around 500 meters at intervals of 30 meters or less allowing a very detailed look at the vertical and spatial structure. Going through the process of gaining access to this data opened the door to other mesoscale data sites operated by state agencies across northern Illinois.