Model Accuracy was onboard the 2nd place overall finisher in the 2021 TransPac. 

Check out the newsletter debrief below!

https://mailchi.mp/f9cfd60549ec/transpac2021-chris 

 

=21pxWhich GRIB file do you usually trust the most when sailing offshore southern California?
Check out a recent analysis from a customer's offshore training session. The boat shared their analysis of weather forecast GRIB files on their journey from San Diego out to the synoptic pressure and back over a day and half of sailing approximately 300 miles.

        Model Accuracy is a critical piece of onboard software which can help navigators learn which weather forecast model is performing the most accurate, and provide calibration trends withinin your routing software. Knowing the trend errors are critical to apply within Expedition routing software to ensure you are always routing with the most accurate GRIB file possible!
        This analysis can be done real time while onboard to make better navigating decisions off the best performing weather forecast GRIB file!Model Accuracy % - the closer the performance line of the GRIB to the top of the graph the better. Here you can see none of the GRIBS correctly forecast the wind shadow behind the Island San Clemente, and once out into the synoptic pressure the SailDocs GFS was the clear winner!TWS Knots Plot - The thin black line is the boat instrument data for TWS. You can see only the HRRR correctly forecast the lee of San Clemente Island, but then the HRRR over predicted the TWS forecast once out in the significant pressure of the syncopic breeze. Once offshore, the GFS and NAM were the most accurate!TWS Error Knots - here one can see the NAM Model did the best job with a near zero error in its prediction once the boat entered the synoptic around 1000Z UTC. [img]https://mcuserc

Version 1.500 has been released! Check out what's new below:

https://mailchi.mp/c3c629c97a99/1point5

Improved User Interface and navigator selection!

We had a very busy winter of working with top level navigators and capitalizing on their feedback. The result is an even more friendly user interface with a host of new and improved features.

List of version 1.500 improvements:

• Added menu to change the time zone so the result reports and graphs can match your local time. *This impacts only how time is displayed as all data must still be in UTC.
• Added features to extract anemometer height and wind type from those log sources that provide that information (otherwise the value entered by the user in the GUI are used). This is mostly for AuBOM and NOAA weather buoy reporting stations. 
• Added an option to enable/disable scaling of TWS. If enabled, both the log TWS and the model TWS are scaled to the specified height (which has no impact on the log TWS if that is the same as the log height, e.g., a boat anemometer). If disabled the theoretical scaling factor is reported in the end result analysis.
• Added ability to analyze multiple log sources. Very helpful with reviewing multiple buoy data at once. 
• Added selection of which logs and which models to analyse. The user can choose after the data has already been prepared. 
• Improved recommendation of start and end analysis times to ensure your GRIBS are always properly paired with your logs. 
• Improved reporting of progress. The progress bar updates more as the software is computing. 
• Improved caching of prepared data to make subsequent analyses much faster.
• Various incremental improvements, many in the user interface. 

https://mailchi.mp/48834996f6b3/northhead
Navigators in Australia, we are so sorry to hear about the cancellation of this year's Sydney-Hobart race.
      In the spirit of still being in "race-mode" we went ahead and ran an analysis of the Predict Wind GFS and EC forecasts against the North Head weather observation station near Sydney. 
       Model Accuracy has a "pre-race" function which allows GRIB files to be compared to weather stations, and we now support the file format of the AuBOM (Australian Bureau of Meteorology) wind observations. This means you can download the weather station log data and use that as the truth against your GRIB file forecasts to see what is the accuracy trend in your area. 
      Take a look below to learn the expect forecast errors are before one ever even heads out to sea!

We compared 2 days of weather instrument data to the PW EC and PW GFS forecasts. We found a significant trend of the TWS forecast being far above what actually occurred, by approximately 7 knots! Check out at the Model Accuracy error graph below of the GRIB forecasts plotted against the observed weather data from North Head weather station.  

Below is the "Results" page of the mathematical and statistical analysis of the forecasts compared to the truth data of the observation station. Both GRIBs were actually similar in their errors of the TWS being 7 knots above and 17-20 degrees right of the observations. Model Accuracy provides exact calibration values to be entered into Expedition Sailing Software for optimal routing of a "corrected" GRIB file.  

Here you can see the TWD error plot. A "perfect" forecast wou

Enjoy our academic paper submission to the INNOVSAIL International Conference!

We were lucky enough to present our software and purpose of its application alongside some very talented and interesting individuals, teams, and companies!

https://static1.squarespace.com/static/542eec32e4b0fc2c609ef58b/t/5fdf4b09fb7de13a646a9280/1608469259645/ModelAccuracy+-+Statistical+Analysis+of+Sailing+Forecasts+-+INNOVSAIL+International+Conference.pdf