I’ve lived in Chicago, Denver, San Francisco and San Diego. Maybe with the exception of Denver, locals seem to be of the belief that accurate weather forecasting in their cities/areas is more difficult than in others. Personally, I think forecasts are more accurate in southern California than in the Chicago area. Of course, the tools forecasters had when I lived there aren’t what they are today. Anyway, is there any evidence (other than anecdotal) that forecasting in certain geographical areas is more or less accurate and/or difficult than others?
Weather forecasting is mostly a guess. They take the current conditions, all of them, and compare them to past conditions and make their best educated prediction (guess) that what happened before (when the same or similar conditions existed) will happen again.
In areas where there are a lot of wildly varying factors the weather that actually occurs can be much different than what was predicted.
Here at the mouth of the Chesapeake it happens all the time. Schools have closed the night before a predicted wallop of a snow is predicted and the next day turns out sunny with not a drop of anything.
If you live in San Diego you don’t have weather - just climate. 
I live in the Bay Area now, but I used to live in NJ. The difficulty of forecasting here is that we have lots of microclimates, so the forecast for SF proper has not much to do with the East Bay or Gilroy. I’d suspect though that in general it is easier than in the East, since the weather patterns are mostly moving in from the Pacific, and there aren’t the chances for change you get when your weather is being affected by the multiple fronts and things you get over land. In general I’ve found the predictions to be better out here than back East.
If by “guess”, you mean some serious computer number crunching, you’d be right. Here’s a USA Today description on weather modeling, along with some links to some good sites.
Compare weather forecasts now against 10 years ago. Or 20 years. Or 30 years. When I was a child, you could ignore 3 day forecasts as being a joke. Now? The long range 5-10 forecasts have a good shot at being accurate.
Forget mobile phones or the internet. The serious improvement in weather forecasts over my lifetime, due mainly to an increase in the number of weather stations and computing power, still manages to flabbergast me.
WAG:
When most people think weather forecasting in their city is harder than others, it’s often a serious case of selection bias. I could see how maybe being on the Pacific coast could moderate things, but I haven’t seen proof of it, either.
There’s an old joke that I’ve heard in every part of the country (except San Diego, actually): Want the weather to change? Wait 10 minutes. I’ve heard that in Seattle, Houston, New York, Chicago, Detroit, Atlanta, Memphis, Baltimore, and I’ve no doubt it’s told elsewhere. I’ve never found it to be more true in any of those places than any other (except again, San Diego).
Here’s an interesting website that gives some numbers for accuracy percentage (based on its own formula) for different cities.
According to the numbers it provides, of the 4 cities you mention (Denver, San Francisco, Chicago, San Diego), the weather forecasting is least accurate in Denver and most accurate San Diego. That’s pretty funny that the folks in Denver are the ones who complained the least.
WAG again, but maybe being just east of the Rockies makes forecasting more difficult.
For what it’s worth, apparently forecasting the mouth of Chesapeake Bay (as measured at Cape Charles, VA) is about on par with Chicago and more accurate than Denver.
Beat me to it. The statistical models are far, far better than they were a few decades ago.
But, even so, they’re not perfect. They still have a hard time handling weather events on a micro scale – for example, getting the “bulls-eye” right on where the heaviest snowfall will fall, or whether a storm will produce rain or snow in an area, if the temperatures are right around freezing.
And, of course, our memories are selective…we remember the (relatively few) events when the forecast was wildly inaccurate, and don’t remember all the days when it was pretty close, if not right on target.
Finally, when it comes to forecasting major severe weather, it does seem to me that some meteorologists are likely to err on the side of forecasting the worst – my supposition is that they would rather be accused of “crying wolf” and predicting a blizzard when we get a dusting, than the reverse, and leaving their viewers / readers unprepared for bad weather.
Many is the time when the So Cal forecast is 7 suns in a row on the telly. Not so hard to get it right when it doesn’t change.
It’s told all the time in Texas during certain parts of the year (i.e. right now). I’ve been outside many times to feel an incoming cold front that can drop the temp 30 to 40 degrees (F) in the span of an hour. Springtime weather here in DFW is notoriously volatile.
However, Texas state law decrees that all meterologists get summers off, same as teachers. It’s the same forecast day in and day out anyway (“Hot. Like, really *really *hot, people. Also, it’ll be hot tomorrow. Hot all weekend, too.”) and nobody likes being reminded that it’s 108 F without the heat index factored in yet.
Even in the Bay Area the six day forecast for highs will be
72 - 72 - 73 - 72 - 75 (warming trend) 75.
Around here if you want the weather to change, drive 10 miles.
Twice in the last half-hour my weather radio has gone off with a severe thunderstorm warning. The sky is cloudy and there was about two minutes of light sprinkles, but nothing else. Exactly 5.18 miles from here they’re having golf-ball size hail.
Around here when we talk about forecasting accuracy, what we really mean is how close we’ll happen to be standing to whatever they happen to be forecasting.