I have been weather forecasting as an amateur for many years. In order to do this, I have three weather stations, each of them able to measure the current weather and interpret it in its own way. In addition, I make enormous use of the weather reported on the Internet over a radius of up to about 1500 km and at altitudes from sea level up to about 15,000 m (tropopause).
Considering that all this is done on a couple of ordinary PCs, my forecasts are mostly very accurate for the part of Cyprus where I live (plonk in the middle!). Although Cyprus is a small island, its topography is sufficiently rugged that the weather can be very different from one place to the next; two places, 10 or 15 km apart, may have very different weather from time to time. This makes weather forecasting quite difficult and I would go so far as to say that it is impossible to be accurate 100% of the time. In reality, I would guess that my own forecasts are better than 90 to 95% accurate for the region where I live and for, say, 20 km or more round it and perhaps even 50 km on a good day! As we are in the centre of the island, I guess that we do not do too badly at all!
Much of the rain that falls on the island is from thunderstorms, which can happen at any time of the year. It is totally impossible, even for professional systems like the Met office, to forecast where a thunderstorm will occur. This problem is exacerbated by the small diameter of many of the storms; I have seen them less than 1 km across. Nobody knows much about how thunderstorms locate themselves, even though the physics of their generation is given; it may be that the presence of a car with a hot engine may trigger one under some circumstances.
My English language forecast, normally updated every three hours, can be found at
http://weather.bnellis.eu/forecast/ and it ain't looking very pretty at the moment!
