“When you can measure what you are speaking about,
and express it in numbers, you know something about it”
– Lord Kelvin
To be clear, Lord Kelvin was known to be pertinently wrong on a lot of things, but we think he nailed this one. The point being that, to us, data quality matters. Spotter is designed from the ground up to make the best possible measurements. We see this development as a continuing journey. So we will continue to post and share updates of data validations by us or others.
To validate data quality we can consider two main methods of testing, namely: collocated statistical comparison, and controlled motion validation. The former compares different instruments measuring the same (in a statistical sense) wave field, the latter compares a prescribed motion to Spotter measurements (deterministic).
Clearly both testing methods are complementary and we use both. In addition, validation can be achieved by comparing data to known theoretical results. This was successfully done by one of our Beta testers and presented at the annual Marine Energy Technology Symposium (see Raghukumar et al, 2017).
The purpose of validation with collocated instruments is to assess in-situ data quality potential, and validate against other types of instruments that measure the same wave field. Since co-located instruments are separated in space, the wave field they measure is not the same in detail, so a direct deterministic comparison of displacement data is not possible. However, since the wave field statistics vary slowly in space, statistical estimates of the two measurements can be directly compared. This is what we refer to as statistical validation.
In April 2017 we deployed a Spotter at our testsite in Half Moon Bay (see figure 1) alongside a Datawell DWG-4 for nearly two weeks. During this period we experienced a range of conditions, ranging from low-energy swell, to energetic locally-generated wind sea, and even a brief period of fetch-limited wave growth (offshore wind). Figure 1 (below) shows a comparison of spectral variables from data recorded on 4/28/17 14:00 UTC in about 40m water depth offshore of Half Moon Bay, and just north of Mavericks. This data comparison and summary of principal results of the validation study can be found in the blog Data Validation: Spotter at Mavericks.
To validate the motion sensing at a more detailed (deterministic) level, we compare Spotter observations against prescribed motion of a motion stand (see figure 2). Our motion-controlled validation stand can run any 2D motion pattern (constrained only by its physical and mechanical limits), so we can test both periodic and random waves (e.g. Jonswap spectrum). Controlled motion validation of Spotter data is discussed in the blog Data Validation: Spotter takes a stand.