Data analysis
We turn raw measurements into clear statistics, trends and findings your engineers and planners can act on.
Measurements become useful when they answer a question. How high do waves get at this site? How often can a vessel work? Which direction do storms come from? Geomarins analyses metocean data and turns it into clear statistics, charts and findings that engineers, planners and managers can act on.
Typical analyses
- Summary statistics: means, maxima and percentiles by month and season
- Joint distributions such as wave height against period, and wind speed against direction
- Wave, wind and current roses showing direction and strength
- Persistence and workability: how often conditions stay within a vessel or crane limit, and for how long
- Weather window statistics for planning marine operations
- Extreme value analysis to estimate 1, 10, 50 and 100-year conditions
- Tidal analysis separating tide from surge and residual currents
- Comparison of measured data with model or forecast data
Working with real data
Measured data has gaps, seasonal bias and short records. A two-year buoy record will not contain the 100-year storm. We say how much confidence each result deserves, and where measured data is too short we combine it with longer model hindcasts, using the measurements to check and correct the model.
Clear results
Each analysis is delivered with plain-language findings alongside the tables and charts, so a project manager can understand the result as well as the engineer. Methods follow recognised industry practice for metocean work, such as ISO 19901-1 and DNV-RP-C205, so results can be used in design.
From analysis to design
When analysis feeds a design or a permit, it often leads into a formal metocean design basis or site conditions report. Our reference document development service takes the analysis into that document.
What you get
- Statistical summaries by month and season
- Wave, wind and current roses
- Workability and weather window statistics
- Extreme value estimates with stated confidence
- Tidal and residual current analysis
- Plain-language findings with every report
Frequently asked questions
How much data do we need for extreme value analysis?
The longer the better. Measured records of a year or two are usually combined with long model hindcasts, with the measurements used to check and calibrate the model.
Can you analyse data you did not collect?
Yes. We quality-check third-party data first, then analyse it in the same way as our own.