2 00:00:08,971 --> 00:00:13,339 In order to simulate the climate we have to simulate the 3 00:00:13,339 --> 00:00:18,418 processes by which this heat gets carried from the equator to the pole. 4 00:00:18,418 --> 00:00:20,070 We have to simulate the weather. 5 00:00:21,740 --> 00:00:28,010 And this is hard because of a phenomenon known as the turbulent cascade. 6 00:00:28,010 --> 00:00:30,400 We're stirring this or 7 00:00:30,400 --> 00:00:34,750 we're driving this with a planetary sized forcing. 8 00:00:34,750 --> 00:00:37,160 The difference between the energy coming in and out at the 9 00:00:37,160 --> 00:00:40,660 equator versus the pull, that's a difference that's as big as 10 00:00:40,660 --> 00:00:43,960 the planet, but the flow that results from that is not 11 00:00:43,960 --> 00:00:48,560 as simple as these nice planet size arrows that I drew. 12 00:00:48,560 --> 00:00:54,570 The flow tends to break up large large eddies, 13 00:00:54,570 --> 00:00:58,200 large motions, and make smaller motions. 14 00:00:58,200 --> 00:00:59,840 It's called the turbulent cascade. 15 00:00:59,840 --> 00:01:04,810 A planetary sized forcing leads to large scale motions which then break up 16 00:01:04,810 --> 00:01:10,270 and break up until, finally, they are the size of the 17 00:01:10,270 --> 00:01:15,350 storm systems that affect us from hour to hour in the weather. 18 00:01:17,910 --> 00:01:22,210 And so this is where climate science starts to get hard. 19 00:01:22,210 --> 00:01:25,680 We simulate the climate of the earth and all of these motions, all of this 20 00:01:25,680 --> 00:01:28,610 weather, on a grid like a checkerboard 21 00:01:28,610 --> 00:01:31,920 in a computer program called a Climate Model. 22 00:01:31,920 --> 00:01:37,620 The surface of the earth is represented by this grid and maybe 23 00:01:37,620 --> 00:01:43,150 you have values for the pressure and the temperature in 24 00:01:43,150 --> 00:01:49,510 the centers of these grid points and maybe you calculate what the winds 25 00:01:49,510 --> 00:01:55,170 should be, the flows in the cells, the, the cell walls between these grid points. 26 00:01:56,310 --> 00:02:00,900 The problem is that we would like to have as much detail as we 27 00:02:00,900 --> 00:02:04,629 can have in our simulations of the weather so we can get things right. 28 00:02:06,520 --> 00:02:08,780 We want to have enough grid points 29 00:02:08,780 --> 00:02:11,860 so that we can adequately see, so that the model 30 00:02:11,860 --> 00:02:16,300 can adequately express the smallest scales of the turbulent cascade. 31 00:02:17,450 --> 00:02:21,750 And the problem is that if we decide to double the 32 00:02:21,750 --> 00:02:25,670 amount of detail in a model, for example, that would mean having 33 00:02:25,670 --> 00:02:29,620 twice as many grid points in the horizontal and twice as many 34 00:02:29,620 --> 00:02:34,030 grid points in latitude, longitude and then also 35 00:02:34,030 --> 00:02:37,610 in the vertical having twice as many grid points. 36 00:02:37,610 --> 00:02:42,960 And so and then, the model steps forward in time and if we 37 00:02:42,960 --> 00:02:48,250 make these boxes smaller it turns out that we have to take smaller steps in time. 38 00:02:48,250 --> 00:02:52,520 A normal climate model might take a time step of 20 minutes, for example. 39 00:02:52,520 --> 00:02:55,180 And if we made it twice as many grid points 40 00:02:55,180 --> 00:02:57,440 maybe we could only take ten minutes per time step. 41 00:02:57,440 --> 00:02:59,610 We'd have to take twice as many steps 42 00:02:59,610 --> 00:03:02,910 to do a day or a year or something like that. 43 00:03:02,910 --> 00:03:06,630 It turns out it's a very steep computational curve. 44 00:03:06,630 --> 00:03:11,630 If we want to double the amount of detail we have to do 16 times more math. 45 00:03:11,630 --> 00:03:15,360 And so it gets much, much harder and much, much slower. 46 00:03:15,360 --> 00:03:19,014 Computers are getting faster but climate scientists are going to leave 47 00:03:19,014 --> 00:03:21,948 the computer room wanting faster computers for a very, very long time.