Operational Filter Design for Non-Contact Water Level Sensors Deployed in an Open Ocean Environment: U.S. Army Corps of Engineers Field Research Facility, Duck, NC July 4 – July 25, 2008 Test Period.

Abstract

Test comparisons of four microwave water level sensors and one laser device for sensing water level were conducted at selected sites on the Atlantic and Pacific coasts as well as Lake Michigan. The microwave and laser devices being evaluated are ‘open-air’, non-contact water level sensors that differ from other sensors in having no stilling well, wave guide or other isolating structure. In this configuration, 1-Hz water level series were seen to contain considerable noise as well as variance at both tidal and non-tidal frequencies, including unrestricted wind-wave variance at higher frequencies. NOS applications related to the astronomical tide (times and heights of high and low water, duration of rise and fall, tidal datum determinations, tidal harmonic analysis, tidal prediction) require effective low-pass filtering to remove unwanted variance as well as noise. A primary filter is required to produce noise-free data at regular intervals that will enable the performance of one sensor to be compared with another - a major objective of the testing. An advanced filter, the Butterworth Infinite Impulse Response (IIR) filter was used for this purpose in combination with a MATLAB forward-backward filtering function that corrects the phase distortion commonly associated with IIR filters. Low RMS difference values for paired series of filtered, zero-mean water level attest to adequate precision among sensors but accuracy limitations reflect the differences noted in series mean water level. This finding may place a premium on sensors producing the least number of outliers.