Mobile Platform Data Quality from Original Equipment Manufacturer (OEM) Sensors on Commercial Vehicles

The National Oceanic and Atmospheric Administration (NOAA) conducts a National Mesonet Program (NMP) that extends beyond the reach of traditional surface observation of weather data, which is highly dependent on Automated Surface Observing Systems (ASOS) at airport locations. Thousands of non-federal weather observing stations record quality environmental observations that are assimilated into NOAA’s operational environment.  Mobile platforms that obtain and report meteorological data further fill in significant spatial gaps in the NMP. The Mobile Platform Environmental Data (MoPED) system, developed and operated by GST acquires environmental-conditions measurements from commercial fleet providers.   MoPED processes the measurement data sets and provides them to the National Weather Service’s Meteorological Assimilation Data Ingest System (MADIS), which is the gateway for non-federal observations into NOAA’s operational environment for further downstream use in operations and numerical models.

In addition to formatting the vehicle measurements for ingest by MADIS, the MoPED system performs routine quality assessments of the vehicles by processing vehicle measurements comparing MoPED vehicle observations to observations produced by ASOS stations when the mobile platforms are in sufficiently close proximity (in space and time). GST has developed a set of quality control algorithms that are applied to vehicle observation datasets to ‘qualify’ them for dissemination to MADIS.  Mobile platforms that do not pass the qualification are removed from the disseminated data stream.

The commercial vehicles that are participating in the NMP have a third-party suite of sensors for ambient air temperature (AT), relative humidity, light level, precipitation intensity, atmospheric pressure, and ozone.  Each commercial vehicle is also equipped with a road temperature sensor for pavement condition.

In addition to third-party sensor packages, vehicles can also generate meteorological data from original equipment manufacturer (OEM) sensors that conform to the Society of Automotive Engineers’ J1939 standard for onboard vehicle networking. These sensors, which can be referred to as OEM J-data sensors, measure a number of parameters associated with emissions control and engine performance, including ambient air temperature and atmospheric pressure. Time-tagged and transmitted to the MoPED system, these OEM J-data measurements could be a tremendous source of weather data for NOAA if the J-data can be extracted and communicated real-time from the vehicles. GST is working with fleet (who have OEM J-data available) to determine suitability for NOAA as a real-time data ingest to MADIS.  GST has developed a methodology for assessing the suitability of classes of vehicles and sensors for inclusion in the MoPED dissemination, for the continued assessment of individual vehicles once their class has been accepted into the MoPED, and for identifying corrective measures (such as adjusting measurements to correct for individual sensor offsets) to ensure the overall quality of the data provided to MADIS. A by-product of that methodology is a multi-component model for sources of errors in mobile meteorological data measurements. In this presentation we describe this error model, and provide examples of GST studies of several different candidate OEM J-data vehicle fleets, in which we measured and/or compensated for various components of the error model.