Multi-sensor evaluation of microwave water level measurement error

Abstract

A microwave radar water level sensor, the Design Analysis Waterlog H-3611 has recently entered service at tide stations operated by the National Oceanic and Atmospheric Administration (NOAA), National Ocean Service (NOS), Center for Operational Oceanographic Products and Services (CO-OPS) as part of the National Water Level Observation Network (NWLON). The microwave water level sensor combines high accuracy with low sensitivity to variations in air temperature and humidity but differs from other water level sensors in utilizing an unconfined radar beam aimed vertically downward to the water surface. Many potential benefits of using microwave radar sensors for short-term flood advisories and long-term sea level monitoring have been identified by several organizations throughout the ocean observing community. The most notable advantage of radar sensors is their ability to measure water level remotely with no parts directly in contact with the water column. Water level measurement stations that employ remote radar sensor technology will avoid many problems typical of long-term subsurface ocean sensors including biological fouling and corrosion. Remote sensing also results in a significant reduction in system hardware components and overall installation and maintenance requirements. Results from a series of laboratory and field tests conducted by CO-OPS over the last few years have led to operational use of Waterlog radar units in certain specific applications. It is acknowledged, however, that most test data collected and analyzed to date for the purpose of assessing the sensor's capabilities have focused on enclosed coastal regions with limited fetch and a low-wave environment (average significant wave height nominally less than 1 m). Although these test results are relevant to many CO-OPS applications of interest, including the majority of NWLON stations located in low-wave environments, uncertainty remained after the initial test phase and led to a new outlook toward sensor performance in open ocean environments that experience significant wave heights frequently in excess of 1 m. Additional testing under these conditions has presented a greater challenge as both the test and reference NWLON sensors are likely to encounter limitations in the presence of large waves. Analytical tools including spectral analysis of sensor output signals and an evaluation of sample statistics were needed to better understand the nature of these limitations. In order to address the remaining uncertainty in sensor performance capability, a multi-sensor test deployment was recently conducted at an open-ocean test site, the U.S. Army Corps of Engineers Field Research Facility at Duck, NC (Duck FRF). The resulting multi-sensor data set has allowed sensor measurement error to be estimated for the first time from measurement residuals about an ensemble average series. Comparing ensemble-based error estimates with corresponding Duck FRF nearshore wave data averaged over selected measurement periods shows that sensor error increases, as expected, with increasing significant wave height. However, the RMS residual error obtained is relatively small for low to moderate waves (Hmo ≤ 1.5 m); e.g., approximately ± 1.5 cm for an individual 6-minute water level measurement over a two-day period with 5-11 s waves averaging 1.14 m in height.