loading page

Ecological Sensor Data Quality Assessed Using Observational Data and Combined Uncertainties
  • +1
  • Guy Litt,
  • Janae Csavina,
  • Joshua Roberti,
  • Jesse Vance
Guy Litt
Battelle Memorial Institute

Corresponding Author:[email protected]

Author Profile
Janae Csavina
National Ecological Observatory Network
Author Profile
Joshua Roberti
National Ecological Observatory Network
Author Profile
Jesse Vance
National Ecological Observatory Network
Author Profile

Abstract

Delivering long-term, high quality environmental sensor data spanning the continent is a primary goal in the National Ecological Observatory Network’s (NEON) Instrumented Systems (IS) group. Some independent observations collected by NEON’s Observation System (OS) measure similar data at the same location and time as the in-situ sensors. Coinciding IS and OS measurements facilitate supplementary data quality assessments by vetting IS sensor data (e.g. aquatic pH probe) against corresponding OS data (e.g. water grab sample analyzed in a lab for pH). To assess whether IS data agree with OS measurements, we use uncertainty as a tool to understand data quality. The uncertainty between NEON IS and OS data follow analytical (e.g. summation in quadrature) or numerical (e.g. Monte Carlo) approaches depending on the complexity of the IS-OS comparison algorithms. NEON calculates the IS-OS uncertainties, and applies the expanded uncertainty as control limits for acceptable IS-OS data comparisons. IS-OS comparisons falling outside the uncertainty-based control limits help to (i) explore unaccounted uncertainty in the IS and OS data, and (ii) address issues in the data or sample collection process as ongoing continuous improvement strategies.