Automated software is a vital tool for quality assurance (QA) of the more than 1 million observations measured daily by the Oklahoma Mesonet. However, even the most carefully designed automated QA routines will miss some erroneous observations. Likewise, some of nature's most interesting meteorological phenomena produce data that fail many automated tests.

The QA meteorologist employs a multitude of manual techniques to complement automated QA. These techniques include: evaluation of the automated QA results on a daily basis to investigate suspicious data, comparison of rainfall observations and radar estimates after each rain event, analysis of monthly-averaged data to detect sensor drift, examination of wind speed data at various heights to detect starting threshold problems, and double mass analyses to check for data consistency. Even more importantly, the QA meteorologist traces the true start time of each problem so that appropriate data can be manually flagged. In addition, the QA meteorologist is responsible for communicating problems to, and coordinating with, appropriate field technicians to ensure proper resolution.

The role of QA meteorologists is also essential to climate networks. Over the past several years, QA meteorologists at the Oklahoma Climatological Survey (OCS) have methodically investigated cooperative observer data for Oklahoma. Manual investigation revealed more than 2000 observations that had been checked by automated routines but turned out to be in error when compared to original records or data from neighboring stations. A daily Top-20 list is one tool used by OCS to identify outliers for further review by the QA meteorologist.