Measurement, Analysis and correlation of Meterological Data

Abstract

Ideally, I would like to have made the study for Kalamazoo. The problem was, a large amount of past data was needed for analysis.If I were to measure data directly, analysis could not begin until the quarter was nearly over. Since the U. S. Weather Service keeps accurate records, I decided to do the study for Grand Rapids, the closest weather service station. To make a complete study, I needed more than just the Grand Rapids data. Data from other stations should prove useful in a predictive scheme. So my proposal became this: will any combination of data taken simultaneously in Grand Rapids and any weather service station within four hundred miles of Grand Rapids yield a device which can accurately predict whether precipitation will occur in the twenty-four hour period immediately following the observation? I also hoped to try something more than just grind out a predictive scheme "by the book," as forecasters have done for years. One would think that with a large volume of data and a computer capable of handling it, it would be possible to come up with a reliable predictive scheme.