Analysis

Updated algorithm improves accuracy of Spotter’s wind velocity measurements

Sofar Ocean

The updated algorithm increases the reliability and accuracy of Spotter’s wind velocity measurements.

Spotter’s estimates of wind velocity are now more accurate and reliable

Sofar’s Spotter buoy collects and transmits ocean data in real-time. The buoy is outfitted with a variety of measurement tools that make direct observations of ocean conditions at the sea surface. Barometric pressure, for example, is recorded using an onboard barometer.

Not all observations reported by Spotters are direct measurements. This ensures that the device can deliver as many ocean insights as possible, without becoming overloaded with specialized and costly measurement devices. Observations that are not made directly are derived:

Spotter makes an observation of a variable that can be directly measured.
The observation is processed through an algorithm.
The algorithm yields a contemporaneous measurement of a variable that can otherwise not be directly measured.

Spotter uses this approach to measure wind velocity. Because the device does not have a dedicated wind sensor onboard and given the strongly coupled relationship between winds and waves, wind velocity is derived from the observed wave spectrum. This is a logical approach, but, like any inference, is susceptible to inaccuracy because several approximations must be made to produce the estimate.

Recently, Tomoya Shimura and his fellow researchers at the University of Kyoto contacted Sofar with a suggested update to the wind velocity algorithm. The change, which was rigorously tested by our team, greatly improves the reliability of Spotter’s estimates of wind velocity, particularly at high speeds. The updated wind velocity algorithm is now available to all Spotter 3 customers via the latest firmware upgrade.

Satellite altimeter data validates updated wind velocity algorithm

To assess the accuracy of the updated algorithm and the original algorithm, we compared the wind velocity magnitude estimates at 10 meters (U₁₀) made by the global grid of Spotter buoys to the estimates of U₁₀ made by altimeters on government-operated satellites. Altimeter observations, which are made as satellites circle Earth, are widely used by forecasting centers. As the satellites sample swaths of ocean that are populated by free-drifting Spotters, we can compare the altimeter and buoys’ estimates of wind velocity in the areas of overlap.

In Figure 1, we compare the collocated observations of wind velocity made by satellite altimeters and our global grid of Spotter buoys from January to December 2022 using the updated (left) and original (right) wind algorithm.

**Figure 1.** Results from comparisons of wind velocity magnitude estimates at 10 meters (U10) as calculated by the updated (Figure 1a) and original (Figure 1b) wind algorithms onboard grid Spotters, and satellite altimeters from January to December 2022. Satellite altimeter values are shown on the x-axis and Spotter values are shown on the y-axis. Each point represents an instance where a satellite measurement and a Spotter measurement were near each other in time and space. An additional condition controlling for significant wave height was implemented to ensure that the satellite and Spotter were measuring conditions in similar sea states. The color scale for the points represents the density of observations. Warm colored regions indicate the most commonly measured wind speed values (around 5 to 9 m/s in both figures). Cool colored regions indicate wind speeds less commonly observed by the instruments included in this analysis. Values for root-mean-square error and bias are 1.41 m/s and 0.15 m/s respectively for the updated algorithm and 1.73 m/s and 0.73 m/s for the original algorithm. Satellite altimeter measurements of wind speed are also derived and their measurements are subject to their own observational errors.

When using the original wind algorithm (Figure 1b, right), we observe that:

Outside of the 5 to 10 m/s range, Spotters consistently underestimate wind speed relative to the satellite altimeters’ observations.
Spotters fail to capture higher values of wind speed, especially above a threshold of about 18 m/s.

When using the updated wind algorithm (Figure 1a, left), we observe that:

The estimates of wind velocity made by grid Spotters and satellite altimeters are more closely aligned across a larger range of wind speed values.
The range of wind speeds estimated by the grid Spotters are more consistent with the range estimated by the satellite altimeters.
At high wind speeds, the Spotter estimates align more closely to the satellite altimeters’ estimates.

Overall, the updated wind algorithm significantly improves the reliability of the Spotter wind velocity estimates when compared to estimates made by satellite altimeters.

How to access the updated wind velocity algorithm

To ensure that your Spotter 3 is using the updated wind algorithm, please install the latest firmware update or reach out to customer support at support@sofarocean.com.

Special thanks to Tomoya Shimura and his fellow researchers at the University of Kyoto —Nobuhito Mori, Yasuyuki Baba, and Takuya Miyashita — for making this update possible. We encourage you to read their full article discussing the wind velocity algorithm in the Journal of Geophysical Research: Oceans. The Sofar team encourages and appreciates feedback. If you have a suggestion for a future update to Spotter or Smart Mooring, please send an email to support@sofarocean.com.

--
To receive the latest posts about marine sensing devices in your inbox, subscribe to our blog. To speak with a Sofar representative, schedule a Spotter smart buoy demo or Data Services demo.

Sofar in the News