The general caveat in forecasting services is that they are totally dependent on input data. Large stretches of the globe are virtually without any surface or air observations. Services in those regions are not as good as services in observation-rich areas. Satellite data improves this situation somewhat, but not entirely. I don't know the observation network in the area she's in, so I won't comment on that.


That being said, the quality of top-of-the-line weather and wave forecasts is really very good. I've been a weather forecaster for 4 years doing pin-point forecasting for the offshore industry, so this is something I know. The main trouble is the cost, not the quality of data.


Obstacle 1: You need computer and data communication equipment that will be reliable on a tiny sailboat. Those are expensive.

Obstacle 2: The results from #1 will have very limited bandwidth. You need a service provider that will feed you high-quality forecast data in a very compact form, suitable for download equipment such as a sat.phone modem for display on your on-board computer - and provides that service at a price you can afford.

All this exist, but quality equipment and services within those constraints will cost you. Dearly.


The computer-and-meteorology savvy can hack together their own forecasting service, extracting data from public available sources and assemble their own collection of products and data. All for free, except for the cost of data transfer. It's not hard if you know what you're looking for. But such "services" will be fragile - what if your techie friend at land has a flu the day you're in the middle of a storm? Or if the original data sources changes their format so the program you've ran for 6 months suddenly stops working? Surfing the web for weather maps on a iridium data modem on a small boat in high seas is NOT a viable plan B.


I know data from many American global models are available in the public domain, but I have never used them, so I can't vouch for their quality, availability or if they provide sufficient detail. Generally speaking, I have no reason to believe they won't do a good job for the purpose at hand, it's just that I haven't checked.


My favourite weather forecasting model is ECMWF weather + wave models, which are global high resolution, so location doesn't really matter. Trouble is, that model is being run by European contries for Medium range Weather Forecasting. For political and economic reasons, access to ECMWF data for short range (1-3 days) is somewhat complicated and/or expensive. Often, a limited area (regional) model is preffered if available, not really because you can provode more detail but because you can update it with fresh observations a lot quicker than the ECMWF model.


EDIT: Obviously, as James_Van_Artsdalen states below, the simplest solution is just to have someone at home to volunteer for daily studying the availabe info and communicate their recommendations.