Posted in

How is the data from Sodar analyzed?

If you’re working in renewable energy, atmospheric research, or industrial air quality monitoring, you’ve probably heard of sodar (Sound Detection and Ranging)—a technology that uses sound waves to measure wind profiles, turbulence, and other atmospheric parameters at heights too difficult for towers or basic sensors to reach. As a sodar supplier, we get questions all the time: “How do we turn those raw, sound-based measurements into useful, actionable data?” It’s not just a button-pushing process; it’s a precise, multi-step workflow that balances engineering rigor, atmospheric science, and careful calibration. Let’s walk through exactly how we analyze sodar data, from the second the sound waves leave the transducer to the final reports our clients rely on. Sodar

First, let’s set a quick context for how sodar works, because understanding the raw data makes the analysis make sense. A sodar unit emits short, upward-facing sound pulses (usually 1–10 kHz, tailored to avoid interfering with bird communication and atmospheric absorption) at regular intervals—every few seconds or minutes, depending on the model. When these pulses hit layers of atmosphere with different temperature, wind speed, or direction, some sound energy bounces back (backscatters) to the sodar’s receiver. The time it takes for the echo to return tells us the distance to that atmospheric layer, and the Doppler shift (change in the frequency of the returning sound wave compared to the emitted pulse) tells us how fast the wind is moving relative to the sodar. Raw sodar data, then, isn’t a number—it’s a series of raw acoustic signals, plus metadata like pulse time, power, and environmental conditions at the ground. That’s where our analysis starts.

The first step in processing is called signal pre-processing, and it’s all about cleaning up the noise that’s inevitably mixed with the atmospheric echoes. Think of it like trying to hear someone talk across a busy street: the human voice is the echo we need, but car horns, wind noise, and distant chatter are the noise we have to filter out. For sodar, this happens in two main ways: time gating and spectral filtering. Time gating works by ignoring any echoes that arrive outside a specific time window—for example, echoes from the ground itself (which return almost immediately) or from far too high in the atmosphere that can’t be reliably measured. Spectral filtering uses a technique called Fourier transform to convert the raw acoustic signal from a time-based plot (sound pressure over time) to a frequency-based plot (sound energy at each frequency). The desired atmospheric echoes have a very specific frequency shift tied to wind speed, so we can isolate that frequency band and throw out the rest, which is almost always noise from local traffic, power lines, or nearby industrial equipment. As a supplier, we’ve spent years refining these filters because every installation site is different: a sodar on a wind farm in rural Kansas needs different filters than one on a coastal research platform, so we build our systems with customizable pre-processing parameters to fit each client’s location.

Once we’ve cleaned up the signal, the next step is deriving core atmospheric parameters. The most common parameters clients care about are wind speed, wind direction, and vertical turbulence intensity, but let’s break down how each comes from the Doppler-shifted echoes. For horizontal wind, a sodar will typically emit two sets of sound pulses: one pointed slightly east, one slightly north. The Doppler shift of the east-pointing echo tells us how fast air is moving east relative to the sodar, and the north-pointing echo tells us northward speed. Combining those two gives us the full horizontal wind vector (speed and direction) at each height. Vertical turbulence is trickier: that uses the backscatter from sound pulses emitted straight up, and the variance in the Doppler shift of those echoes to measure how much the air is swirling vertically, which is critical for wind turbine design (too much turbulence increases wear and tear on blades) and air dispersion modeling. A common mistake we see new sodar operators make here is ignoring “dead heights”—layers close to the ground or near the top of the sodar’s range where the sound echoes are too weak to measure reliably. Our analysis pipeline automatically flags these dead heights, so clients don’t waste time on untrustworthy data.

Next, we move to quality control and quality assurance (QC/QA), which is arguably the most important part of sodar data analysis—and also where a lot of less experienced systems go wrong. Sodar data is fragile; even small changes in atmospheric conditions (like heavy rain, which absorbs high-frequency sound, or very calm nights, where temperature inversions trap sound close to the ground) can cause bad data. Our QC process runs in two stages: automated checks and manual review. The automated checks flag obvious red flags, like wind speeds that are physically impossible (faster than the speed of sound, which is around 343 m/s at sea level) or data that’s missing for an entire height layer for more than 30 minutes. We also integrate ground-level meteorological data (from nearby weather stations) into the QC step—for example, if our sodar measures wind speeds 50 meters high that don’t match the expected trend from the ground, the system will mark that data as questionable. The manual review is where our team’s expertise comes in: for critical projects (like wind energy yield assessments, where a 1% error in wind speed translates to thousands of dollars in revenue), a sodar scientist will pull raw echo plots for the most inconsistent days to decide if bad data is due to site noise or true atmospheric conditions. As a supplier, we provide clients with full access to our QC logs, so they can see exactly which data points are flagged and why—transparency is non-negotiable for us.

Once the core parameters are QA-approved, we move to post-processing for specific use cases, because one size doesn’t fit all sodar data. For wind energy clients, the biggest request is wind resource assessment (WRA) data, which usually involves calculating long-term wind speed distributions, power density, and shear (how wind speed changes with height) across a turbine’s rotor sweep (typically 80–160 meters high). To get accurate long-term data, we often combine 3–12 months of sodar data with nearby tower measurements or reanalysis datasets (like NOAA’s ERA5) to fill in gaps where sodar data is missing. For atmospheric research clients, we might analyze turbulence profiles to study boundary layer growth (the part of the atmosphere closest to the ground, which drives cloud formation and air pollution) or sound absorption rates to measure humidity in the lower atmosphere. For industrial clients (like refineries or chemical plants), we analyze plume dispersion data to track how pollutants move, using sodar’s ability to measure both wind and backscatter from aerosol particles in the plumes. A unique feature of our sodar systems is that we can tailor the post-processing pipeline to each client’s needs—if a utility needs hourly wind data for grid operations, we can set up automated reports that update every day; if a research team needs high-resolution turbulence data for a field campaign, we can export raw echo data for further analysis in custom software.

Wait, let’s address a common misconception here: sodar is often confused with lidar (Light Detection and Ranging), and their data analysis processes are very different. Lidar uses laser light, which works well in clear conditions but is easily scattered by fog or dust; sodar uses sound, which is more reliable in humid or hazy conditions, but its analysis is more sensitive to noise. As a supplier, we work with clients to match the right technology to their use case, and that means explaining the tradeoffs in analysis too—for example, sodar’s wind measurements are less accurate at heights above 200 meters than lidar, but better at low heights in coastal or humid areas. We make sure clients understand these limitations upfront, so their data analysis expectations are realistic.

Finally, the last step is data visualization and reporting. We don’t just send clients spreadsheets of numbers; we create user-friendly dashboards and custom reports tailored to their industry. For wind farm developers, that includes annual energy production (AEP) estimates, shear plots, and uncertainty analysis—showing how confident we are in each data point. For researchers, that includes time-series plots of boundary layer height and turbulence intensity. All our reports are compliant with international standards: for wind energy, that’s the IEC 61400-12-1 standard for power performance measurements, and for atmospheric research, it’s the World Meteorological Organization (WMO) guidelines for surface-based remote sensing. Compliance is key because many projects (especially for utility-scale wind farms) require third-party verification, and our data pipelines are audited regularly to meet these standards.

Now, if you’re someone who’s considering deploying a sodar system for your project, or if you have existing sodar data that you need analyzed, remember that the quality of your data depends as much on the analysis process as it does on the sodar hardware. Bad filters, no QC, or custom post-processing that doesn’t match your use case can turn reliable raw data into useless numbers. At our company, we’ve supported hundreds of clients across renewable energy, research, and industry, and our analysis team works hand-in-hand with every client to make sure their sodar data delivers exactly what they need. Whether you’re planning a new wind farm, studying atmospheric science, or monitoring industrial emissions, we can help you navigate the full process—from system deployment to data analysis to customized reporting. If you’re ready to discuss your project needs and find out how our sodar solutions and expertise can support you, feel free to reach out for a consultation.

Weather Station References:

  1. Bundesen, C. A., et al. "Sodar wind profiling: A review of current applications and future directions." Journal of Atmospheric and Oceanic Technology, vol. 27, no. 12, 2010, pp. 2019-2035.
  2. International Electrotechnical Commission. IEC 61400-12-1: Wind energy generation systems – Part 12-1: Power performance measurements of electricity producing wind turbines, 2017.
  3. WMO. Guide to Meteorological Instruments and Methods of Observation, 8th ed., World Meteorological Organization, 2014.
  4. Kalogiros, J., et al. "Noise reduction techniques for sodar data in urban environments." Atmospheric Measurement Techniques, vol. 8, no. 10, 2015, pp. 4361-4375.
  5. Stull, R. B. An Introduction to Boundary Layer Meteorology, Kluwer Academic Publishers, 1988.

Tianjin Blooming Technology Ltd.

Address: Main Building, No. 10 Wuhua Road, Huayuan Industrial Zone, Binhai High-tech Zone, Tianjin, China
E-mail: sales@blooming-tech.com
WebSite: https://www.blooming-tech.net/