Posted: 27/06/2015 at 13:11
If you put a huge steel bathtub with all sorts of interesting rocks in it as well as a lot of water onto the stove, though, you'll get currents through the rocks and those currents will change with the temperature. The Earth is way more complicated than a bathtub full of rocks and water. We've got reflectivity, clouds, glaciers, deep ocean currents and all that. Different parts can be affected in different ways. If Alaska (which is a very big place) gets 1C warmer and Britain (which is relatively small) gets 4C cooler, that's an increase in the average temperature. It'd be a bugger for both places, with melting permafrost causing problems in Alaska and night frosts through much of the year in Britain, too.
Now lets look at "annual." It's an annual average. We're looking at one temperature for the whole year, not the highest summer temperature, the coldest winter temperature, the temperature at dawn on 21 March or anything specific like that.
If December, January, February and March are 0.1C warmer at night and the rest of the year stays the same, that's an increase in the average, and you'd probably call it pleasant if you noticed. It might result in wasp colonies surviving the winter and growign to insane numbers in some places, or malaria-carrying mosquitoes breeding in new territories or something, but it'd reduce frost deaths in the gardens and cold deaths among the elderly.
If June, July and August are 0.1C warmer during the day, that's an increase and you'd barely notice. Let's call it 1C. Now you're looking at more heat deaths in hot areas and more skin cancer among the fools doing unprotected sunbathing.
If the whole year gets 1C warmer, obviously that's an increase.
If December's 1C colder, January and February are 2C colder, March is 1C colder, May's the same, June's 1C hotter, July's 2C hotter, August's 3C hotter, September's 2C hotter and October and November are the same, that's an increase in the annual average temperature in your location. It comes with all the above problems.
Now combine "global" and "annual" and you get this: if Alaska gets 1C warmer in July, August and September and Britain gets 3C colder in December and March and 5C colder in January and February, that's an increase in the global annual average temperature.
Snow in your garden does not mean there's no global warming.
When it comes to extreme weather, no we can't say "that tornado hit that house because of climate change." No, we can't say "those guys drowned in their attic because of flooding caused by climate change." We don't have another Earth without the climate change. We can't do a comparison. What we can say is that the models all include an increase in the frequency and severity of extreme weather events. Going back to the bathtub on the stove, the higher you turn the heat the more often a big bubble goes "bloip" somewhere on the surface. If you've got a hand over the surface you can't say "that droplet of boiling water hit my hand because the heat was turned up to 95% instead of 94%" but you can predict that you'll get splashed with more droplets of boiling water because of that increase. We can't predict what time it'll start to rain on the opening day of Wimbledon in 2025. We can, however, map rainfall and detect changes in its location, frequency and intensity. If your reservoirs only hold enough water for 30 days, you're fine with getting 4 days' water every 3 days or 6 days' water every 5 days or 20 days' water every 18 days, but getting 70 days' water in 2 days then nothing for the next 60 days is two very big problems: flood and drought. If all the rain falls downstream of the reservoirs, that's a problem too. It's only really useful if it lands upstream of them. You could build a dam across the Nidd gorge at Knaresborough, flood Patel