Browse Prior Art Database

Generating short term rain forecasts from amateur weather station data

IP.com Disclosure Number: IPCOM000236084D
Publication Date: 2014-Apr-04
Document File: 1 page(s) / 35K

Publishing Venue

The IP.com Prior Art Database

Abstract

An algorithm is presented for generating short term rain forecasts from a set of amateur weather stations. The novel feature of the algorithm is that it does not require knowledge of the exact location of the stations used to build the forecast.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 59% of the total text.

Page 01 of 1

Generating short term rain forecasts from amateur weather station data

Current weather forecasts algorithms work by building a complex model of the atmosphere and hence they can only work if the location of each of the weather stations is known with confidence and the sensors are all well calibrated.

In recent years there has been an explosive growth in the number of private

weather stations run by non-experts. It would be advantageous to be able to leverage such data sources, but the stations often don't report their location properly and often the sensors are poorly calibrated.

In this disclosure an algorithm has been developed that can leverage private

weather stations of uncertain location and/or poor calibration to still provide an excellent local forecast for the chances of rain over the next few hours. Such a forecast could be vital, e.g. it could be used in deciding whether or not to leave the roof of a stadium open during a match.

    Instead of building a complex and detailed model of atmospheric conditions the user instead tries to guess which weather stations are less than n hours upwind of the target station (where N is the number of hours for which a rain forecast is

needed).

This algorithm works in three key steps/phases:


Initially examine the daily rainfall measurements over the last month to see which


1.

stations have daily rainfall figures which are closely correlated with the target station (e.g. using pearson correlation coefficient - see http://...