About Weather Data
Weather data describes the atmosphere at a particular time and place. It may include temperature, relative humidity, wind direction, wind speed, barometric pressure, visibility, precipitation, and other factors. There are two basic types of weather observations:
- Surface Observations — Readings taken by an observer from weather instruments which are about five feet off the ground. Readings may be hourly, daily, weekly, or monthly.
- Upper-Air Observations — Taken from a radio transmitter attached to weather sensors. The sensors trail a balloon as it rises from the ground up to 100,000 feet.
Data Rescue Problems and IEDRO’s Solution
Problems with Historic Observations
- Observations — Records of unusual environmental disturbances were recorded in ships’ logs, personal diaries, newspaper accounts, or other temporary media before the invention of modern meteorological instruments. Many times the data was imperfect or estimated. For example, winds at sea may have been described by sea conditions or the billowing of ships’ sails, while winds on land may have been estimated by the movement of trees and other flora.
- Recording Methods — Prior to 1985, weather observations were recorded on paper and transmitted via radio, telephone, and teletype to weather forecasters. Unfortunately, many times transmission difficulties prevented the readings from reaching users. This handicapped scientists wishing to study a particular weather phenomenon, since they did not have a complete set of weather observations archived.
Strip charts were used to record observational data such as temperature, atmospheric pressure, relative humidity, precipitation, stream flow, wind direction and speed, etc.
“The further back you look, the further forward you can see.”
—Winston Churchill
Like the one above, hydro-meteorological “strip charts” look like pen traces on a grid. The charts are mounted on cylinders that rotate at a constant speed; a pen, attached to a mechanical device, records the changes in parameter values over time. Depending on the speed at which the cylinder turns, the charts may represent parameter changes over a 24-hour period or a 7-day week.
The Problem
The process to extract information and digitize the values from strip charts involves manually running a digitizing pen over the chart to trace it and then entering the data on a form by hand. This process takes 15 to 20 minutes of effort per chart.
IEDRO’s Solution
Dr. Ed Root, one of IEDRO’s talented volunteers, is developing a computer program that will copy and digitize a scanned or photographed image of a strip chart. 
The digitized data will be provided in a comma-delimited file or transcribed in a table. The time required for the program to digitize a chart will be less than 5 seconds!
Stay tuned to IEDRO through our website or our free monthly newsletter. Sign up at admin@iedro.org.