GIP Magazine | Geoinformation Processing Department @ ITC

Optimized use of Meteorological Crowd Sourced Data

Overview of the WOW-NL network

Objective:

The key objective of the project is to use machine learning techniques to validate and correct WOW measurements in real-time by integrating high-resolution ancillary (geo)datasets. Ultimately, this will lead to high resolution products and a real-time quality feedback system for WOW-participants. Scientific challenges: • Application and understanding of Machine Learning Analysis (Big Data Analysis). • Integration of large (geo-)datasets (Big Data engineering). • Development of (grid) data quality indicators enabling quality assessment. • Development of tools for real-time validation of third party (WOW) data. • Real-time data integration, processing and validation.

Keywords:

crowdsorucing, weather station, machine learning, big data

Example of a WOW measurement site. This station measures temperature, precipitation, humidity and wind.

Description:

At present there is omnipresent demand for higher spatial and temporal resolution weather and climate data from both society and KNMI operations. Crowd sourced measurements provide opportunities to complement the KNMI network of Automatic Weather Stations (AWS). Since April 2015 KNMI actively collects measurements from weather enthusiast through the crowd sourced WOW platform. The WOW-NL network measures at 311 locations (Fig. 1). Among the crowd sourced weather networks in the Netherlands, WOW is not the largest network, but certainly the best documented and accessible crowd sourced network for KNMI research. In contrast with the KNMI AWS-network, WOW stations are to a large extent located in urban areas and non-standardized locations.

There are several opportunities for projects using crowd sourced data from the WOW platform. Data validation certainly is an important aspect, two previous researches corrected measurements using solar radiation data. The validation can be improved with other external data sources. With the integration of large datasets our goal is to explain the local variation in temperature, precipitation and wind. If you have your own ideas for a project using the crowd sourced data from the WOW platform, please do contact us.