ENANDES+QC: Automated Quality Control Programs for Hourly Climate Observations from Conventional Weather Stations

This project shares a suite of Python-based tools developed under the ENANDES+ initiative to perform automated quality control of hourly climate observations collected from conventional weather stations in Ecuador. The programs cover the following variables: Precipitation (00h, 12h, 18h UTC), Air temperature (dry-bulb) (00h, 12h, 18h UTC), Maximum temperature, Minimum temperature The methodology integrates physical filters, robust statistical techniques (e.g., modified Z-score, empirical percentiles), probabilistic modeling (Gamma distribution), and unsupervised machine learning algorithms such as Isolation Forest and KMeans. These tools follow the recommendations of the World Meteorological Organization (WMO) and the CRC-SAS and aim to ensure reliable and traceable climate data for analysis, modeling, and decision-making.