1 00:00:09,790 --> 00:00:06,850 [Music] 2 00:00:11,200 --> 00:00:09,800 morning everyone my name is Allyson wing 3 00:00:14,199 --> 00:00:11,210 I'm an assistant professor at Florida 4 00:00:16,269 --> 00:00:14,209 State University and most of you don't 5 00:00:18,999 --> 00:00:16,279 know me because I do not study 6 00:00:21,249 --> 00:00:19,009 exoplanets or planetary atmospheres 7 00:00:24,519 --> 00:00:21,259 I studied tropical convection on earth 8 00:00:26,410 --> 00:00:24,529 but I was really excited to be invited 9 00:00:28,990 --> 00:00:26,420 by Ray to come here and speak to you 10 00:00:31,210 --> 00:00:29,000 about some recent work on something 11 00:00:32,679 --> 00:00:31,220 called convective cell bagra Gale be 12 00:00:34,479 --> 00:00:32,689 really interested to hear about and 13 00:00:37,000 --> 00:00:34,489 really interested to talk to you guys 14 00:00:39,700 --> 00:00:37,010 afterwards to hear you think about how 15 00:00:43,479 --> 00:00:39,710 this might be happening or could happen 16 00:00:45,789 --> 00:00:43,489 on exoplanet atmospheres so the kind of 17 00:00:48,910 --> 00:00:45,799 background for this is that in Earth's 18 00:00:51,280 --> 00:00:48,920 atmosphere convection organizes on a 19 00:00:52,840 --> 00:00:51,290 variety of spatial and temporal scale so 20 00:00:55,210 --> 00:00:52,850 it's not just scattered about but it's 21 00:00:57,520 --> 00:00:55,220 often clumped together in some way and 22 00:00:58,960 --> 00:00:57,530 though it's organized on fairly small 23 00:01:00,820 --> 00:00:58,970 scales like in the order of hundred 24 00:01:02,590 --> 00:01:00,830 kilometers or so and things like squall 25 00:01:05,200 --> 00:01:02,600 lines you know groups of thunderstorms 26 00:01:07,690 --> 00:01:05,210 in a line things called mesoscale 27 00:01:10,870 --> 00:01:07,700 convective systems like this one shown 28 00:01:13,030 --> 00:01:10,880 here that's over India but convection is 29 00:01:14,770 --> 00:01:13,040 also organized on larger scales around a 30 00:01:16,900 --> 00:01:14,780 thousand kilometer scale in features 31 00:01:20,170 --> 00:01:16,910 like tropical cyclones hurricanes and 32 00:01:22,630 --> 00:01:20,180 typhoons or equatorial wave so this is 33 00:01:24,550 --> 00:01:22,640 just this band of convection propagating 34 00:01:27,160 --> 00:01:24,560 off of Africa here it's a easterly wave 35 00:01:28,600 --> 00:01:27,170 for example and then we have even have 36 00:01:30,040 --> 00:01:28,610 convection that's organized on the 37 00:01:32,350 --> 00:01:30,050 planetary scale so there's something 38 00:01:36,990 --> 00:01:32,360 called the mad and Julian oscillation 39 00:01:39,820 --> 00:01:37,000 which is a you know thousands of scale 40 00:01:42,520 --> 00:01:39,830 blob of clouds and convection that 41 00:01:44,680 --> 00:01:42,530 slowly moves from the Indian Ocean to 42 00:01:46,690 --> 00:01:44,690 across the Pacific Ocean with a time 43 00:01:48,760 --> 00:01:46,700 scale of 30 to 60 days it's also called 44 00:01:50,770 --> 00:01:48,770 the interest seasonal oscillation so 45 00:01:54,010 --> 00:01:50,780 this organization of convection is 46 00:01:55,719 --> 00:01:54,020 really ubiquitous and all of these sorts 47 00:01:58,000 --> 00:01:55,729 of types of arginase it organization can 48 00:02:00,700 --> 00:01:58,010 be caused by a variety of different 49 00:02:03,490 --> 00:02:00,710 mechanisms and so convection can be 50 00:02:06,610 --> 00:02:03,500 organized by vertical wind shear this is 51 00:02:07,990 --> 00:02:06,620 when winds change with altitude that's 52 00:02:11,020 --> 00:02:08,000 something that helps those squall lines 53 00:02:13,270 --> 00:02:11,030 organize for example convection can be 54 00:02:15,940 --> 00:02:13,280 organized by sea surface temperature 55 00:02:18,250 --> 00:02:15,950 gradients where the chemical clouds and 56 00:02:20,110 --> 00:02:18,260 convection tend to occur where we have 57 00:02:20,650 --> 00:02:20,120 our warmer sea surface temperatures 58 00:02:23,740 --> 00:02:20,660 compared 59 00:02:26,830 --> 00:02:23,750 the cold regions as well as by dynamical 60 00:02:28,930 --> 00:02:26,840 disturbance so things that arise out of 61 00:02:33,100 --> 00:02:28,940 instabilities in the dynamical flow 62 00:02:34,720 --> 00:02:33,110 itself so those are some ways that 63 00:02:36,040 --> 00:02:34,730 convection can organize but today I'm 64 00:02:38,650 --> 00:02:36,050 going to talk about something different 65 00:02:41,470 --> 00:02:38,660 I'm gonna talk about a particular type 66 00:02:44,050 --> 00:02:41,480 of convective organization called self 67 00:02:46,750 --> 00:02:44,060 aggregation but before I get into 68 00:02:49,150 --> 00:02:46,760 explaining what self aggregation is I 69 00:02:51,460 --> 00:02:49,160 want to just give you some background on 70 00:02:52,960 --> 00:02:51,470 atmospheric convection more generally 71 00:02:56,140 --> 00:02:52,970 because I know that we have a bit of a 72 00:02:57,610 --> 00:02:56,150 mixed audience here so to talk about 73 00:02:59,140 --> 00:02:57,620 convection I actually want to take it 74 00:03:01,240 --> 00:02:59,150 even further step back and talk about 75 00:03:04,030 --> 00:03:01,250 radiation which I know that many of you 76 00:03:07,480 --> 00:03:04,040 know quite a lot about so thinking about 77 00:03:08,530 --> 00:03:07,490 the situation on earth what would what 78 00:03:10,660 --> 00:03:08,540 does it look like if we only have 79 00:03:12,610 --> 00:03:10,670 radiative processes going on where we 80 00:03:14,560 --> 00:03:12,620 have shortwave radiation from the Sun 81 00:03:16,960 --> 00:03:14,570 that heats the surface of the earth and 82 00:03:18,910 --> 00:03:16,970 the atmosphere and then the surface in 83 00:03:20,770 --> 00:03:18,920 the atmosphere then cool by emitting 84 00:03:22,930 --> 00:03:20,780 long wave radiation and if this is the 85 00:03:24,810 --> 00:03:22,940 only thing that goes on the planet will 86 00:03:27,040 --> 00:03:24,820 enter a state of radiative equilibrium 87 00:03:28,900 --> 00:03:27,050 which is the equilibrium state of the 88 00:03:32,290 --> 00:03:28,910 combined atmosphere surface system in 89 00:03:34,390 --> 00:03:32,300 the absence of non-rainy reflexes and so 90 00:03:36,580 --> 00:03:34,400 what the act of radiative heating does 91 00:03:37,930 --> 00:03:36,590 is to try to drive the actual state 92 00:03:40,210 --> 00:03:37,940 towards the state of radiative 93 00:03:41,770 --> 00:03:40,220 equilibrium and this has several 94 00:03:44,020 --> 00:03:41,780 important properties its properties of 95 00:03:47,080 --> 00:03:44,030 course depend on the composition of the 96 00:03:48,640 --> 00:03:47,090 atmosphere so when we model this you 97 00:03:51,460 --> 00:03:48,650 have to put in the particular trace gas 98 00:03:53,410 --> 00:03:51,470 profiles that your planet has and one 99 00:03:55,390 --> 00:03:53,420 important thing is a non-local effects 100 00:03:57,340 --> 00:03:55,400 of radiation so changes in one layer 101 00:03:59,710 --> 00:03:57,350 affect others and that turns out to be 102 00:04:02,979 --> 00:03:59,720 important later on for this convective 103 00:04:04,870 --> 00:04:02,989 organization so you can of course do a 104 00:04:06,550 --> 00:04:04,880 radiative transfer calculation to 105 00:04:08,920 --> 00:04:06,560 calculate the profile of temperature you 106 00:04:10,690 --> 00:04:08,930 would get if you put in the particular 107 00:04:12,910 --> 00:04:10,700 parameters for Earth or any other planet 108 00:04:16,360 --> 00:04:12,920 you're interested and so this was first 109 00:04:17,650 --> 00:04:16,370 done you know many many decades ago for 110 00:04:20,229 --> 00:04:17,660 Earth and this is just an example of a 111 00:04:21,789 --> 00:04:20,239 one-dimensional radiative model from an 112 00:04:23,710 --> 00:04:21,799 avi and Strickler and it shows the 113 00:04:25,780 --> 00:04:23,720 temperature profile that you would get 114 00:04:27,940 --> 00:04:25,790 and so there are several features here 115 00:04:29,230 --> 00:04:27,950 that are you know more or less correct 116 00:04:31,719 --> 00:04:29,240 for Earth we have temperature that 117 00:04:33,129 --> 00:04:31,729 decreases with height in the layer close 118 00:04:35,050 --> 00:04:33,139 to the surface and 119 00:04:36,249 --> 00:04:35,060 then after reaching a cold point it 120 00:04:37,209 --> 00:04:36,259 turns around and increases in the 121 00:04:38,980 --> 00:04:37,219 stratosphere this is because the 122 00:04:40,989 --> 00:04:38,990 composition changes here we have heating 123 00:04:44,080 --> 00:04:40,999 from ozone that causes this temperature 124 00:04:45,820 --> 00:04:44,090 increase in the stratosphere but if this 125 00:04:47,379 --> 00:04:45,830 is not quite exactly what the 126 00:04:49,510 --> 00:04:47,389 temperature profile on earth looks like 127 00:04:51,459 --> 00:04:49,520 so what's wrong with this well first of 128 00:04:53,769 --> 00:04:51,469 all the surface is way too hot it's not 129 00:04:56,580 --> 00:04:53,779 340 Kelvin on the surface of the earth 130 00:05:02,830 --> 00:04:59,739 are the tropopause temperature here is 131 00:05:05,830 --> 00:05:02,840 also way too cold it should be closer to 132 00:05:07,629 --> 00:05:05,840 200 or 220 Kelvin and the difference in 133 00:05:09,820 --> 00:05:07,639 temperature the lapse rate how 134 00:05:12,749 --> 00:05:09,830 temperature decreases with height is too 135 00:05:15,550 --> 00:05:12,759 large it falls off much too quickly so 136 00:05:17,110 --> 00:05:15,560 what is missing why why is this not 137 00:05:18,550 --> 00:05:17,120 right why does this not reflect what 138 00:05:21,429 --> 00:05:18,560 things actually look like on earth well 139 00:05:25,839 --> 00:05:21,439 what's missing is convection so it turns 140 00:05:27,640 --> 00:05:25,849 out that this lapse rate that we have in 141 00:05:31,300 --> 00:05:27,650 the troposphere is one that is very 142 00:05:33,390 --> 00:05:31,310 unstable to convection okay so a parcel 143 00:05:37,420 --> 00:05:33,400 of air that rises in the atmosphere 144 00:05:39,189 --> 00:05:37,430 would cool less than what this 145 00:05:40,300 --> 00:05:39,199 environment would would be doing and 146 00:05:41,740 --> 00:05:40,310 therefore it'd be continued to 147 00:05:44,439 --> 00:05:41,750 accelerate and convex 148 00:05:46,029 --> 00:05:44,449 however the stratospheric profile looks 149 00:05:47,559 --> 00:05:46,039 pretty decent and so actually radiative 150 00:05:49,629 --> 00:05:47,569 equilibrium works pretty well for the 151 00:05:51,969 --> 00:05:49,639 stratosphere but in the troposphere this 152 00:05:53,860 --> 00:05:51,979 layer where weather occurs close to the 153 00:05:56,769 --> 00:05:53,870 surface we need to consider convection 154 00:06:00,159 --> 00:05:56,779 so we go back to our little schematic 155 00:06:02,170 --> 00:06:00,169 here this is unstable and so it's not 156 00:06:03,999 --> 00:06:02,180 realized in earth atmosphere and we 157 00:06:05,679 --> 00:06:04,009 begin to have overturning and convection 158 00:06:08,679 --> 00:06:05,689 occur where we have then turbulent 159 00:06:10,719 --> 00:06:08,689 fluxes of sensible heat that transport 160 00:06:12,579 --> 00:06:10,729 heat in the vertical and so here I'll 161 00:06:13,869 --> 00:06:12,589 call this radiative dry convective 162 00:06:16,240 --> 00:06:13,879 equilibrium because we haven't talked 163 00:06:18,329 --> 00:06:16,250 about about moisture yet and what that 164 00:06:20,950 --> 00:06:18,339 is then is a statistical equilibrium 165 00:06:23,230 --> 00:06:20,960 between the net radiative cooling of the 166 00:06:26,679 --> 00:06:23,240 atmosphere and these turbulent fluxes of 167 00:06:29,439 --> 00:06:26,689 heat okay so we can you know take that 168 00:06:31,570 --> 00:06:29,449 into account in into our model and and 169 00:06:33,760 --> 00:06:31,580 model a state of rate of dry convective 170 00:06:36,300 --> 00:06:33,770 equilibrium and this is what we get the 171 00:06:39,700 --> 00:06:36,310 second line here where we allow it to 172 00:06:41,290 --> 00:06:39,710 connect to a neutral state for dry 173 00:06:43,869 --> 00:06:41,300 convection which is called the dry 174 00:06:46,430 --> 00:06:43,879 adiabatic lapse rate so this looks 175 00:06:48,590 --> 00:06:46,440 better than I did before but there still 176 00:06:50,420 --> 00:06:48,600 some problems the surface of the earth 177 00:06:52,880 --> 00:06:50,430 is still a bit too hot 178 00:06:53,540 --> 00:06:52,890 it's like 320 Kelvin or so that's still 179 00:06:56,660 --> 00:06:53,550 too warm 180 00:06:58,850 --> 00:06:56,670 our tropopause is still too cold and our 181 00:07:01,580 --> 00:06:58,860 lapse rate is still a bit too large so 182 00:07:03,740 --> 00:07:01,590 again in dry RCE you ready to come back 183 00:07:05,930 --> 00:07:03,750 to equilibrium the last rate is -10 184 00:07:07,460 --> 00:07:05,940 Kelvin per kilometer but in reality we 185 00:07:09,350 --> 00:07:07,470 know from measurements that temperatures 186 00:07:11,870 --> 00:07:09,360 fall off with Heights at around a rate 187 00:07:13,940 --> 00:07:11,880 of 6 and a half Kelvin per kilometres so 188 00:07:16,130 --> 00:07:13,950 we're still missing something here and 189 00:07:18,470 --> 00:07:16,140 what we're missing is the fact that we 190 00:07:21,080 --> 00:07:18,480 need to conserve moist convection in 191 00:07:23,510 --> 00:07:21,090 Earth's atmosphere and this is because 192 00:07:25,660 --> 00:07:23,520 above a very thin boundary layer near 193 00:07:28,760 --> 00:07:25,670 the surface most atmospheric convection 194 00:07:30,560 --> 00:07:28,770 involves a phase change of water and so 195 00:07:32,930 --> 00:07:30,570 if we imagine a parcel of air that is 196 00:07:34,700 --> 00:07:32,940 rising the atmosphere it's up as it 197 00:07:36,380 --> 00:07:34,710 rises it cools and at some point it 198 00:07:38,480 --> 00:07:36,390 reaches its saturation and vapor 199 00:07:41,870 --> 00:07:38,490 pressure and then condensation occurs 200 00:07:44,420 --> 00:07:41,880 and you have a cloud forming and once 201 00:07:47,570 --> 00:07:44,430 it's saturated it then rises at a 202 00:07:49,220 --> 00:07:47,580 different rate because as this I do have 203 00:07:51,080 --> 00:07:49,230 the condensation in cloud formation here 204 00:07:53,240 --> 00:07:51,090 you're releasing latent heat into that 205 00:07:55,340 --> 00:07:53,250 add some heat back into the system and 206 00:07:58,760 --> 00:07:55,350 you don't therefore cool as quickly as 207 00:08:00,950 --> 00:07:58,770 if you were dry so the picture we have 208 00:08:02,960 --> 00:08:00,960 in the end is this state of radiative 209 00:08:04,610 --> 00:08:02,970 convective equilibrium where again it's 210 00:08:06,200 --> 00:08:04,620 a statistical equilibrium between the 211 00:08:08,020 --> 00:08:06,210 net rate of cooling and turbulent Lexa's 212 00:08:11,300 --> 00:08:08,030 of heat but now you have to consider 213 00:08:13,220 --> 00:08:11,310 latent heat as well evaporation and 214 00:08:14,900 --> 00:08:13,230 condensation heating and we have 215 00:08:18,020 --> 00:08:14,910 transport of heat in the vertical by 216 00:08:19,990 --> 00:08:18,030 these rising air parcels warm air 217 00:08:22,430 --> 00:08:20,000 parcels in in clouds 218 00:08:23,780 --> 00:08:22,440 so that's rate of convective equilibrium 219 00:08:25,670 --> 00:08:23,790 so that's this is how I think about how 220 00:08:27,380 --> 00:08:25,680 convection arises in the Earth's 221 00:08:29,090 --> 00:08:27,390 atmosphere it's because our radio 222 00:08:30,620 --> 00:08:29,100 equilibrium state is unstable and so 223 00:08:32,690 --> 00:08:30,630 therefore it convex and we have in the 224 00:08:35,120 --> 00:08:32,700 tropical atmosphere this occurring you 225 00:08:36,529 --> 00:08:35,130 know everywhere basically so what are 226 00:08:38,600 --> 00:08:36,539 the consequences of the fact that we 227 00:08:41,690 --> 00:08:38,610 have this moist convection occurring in 228 00:08:43,130 --> 00:08:41,700 the atmosphere well if we take our sort 229 00:08:44,990 --> 00:08:43,140 of one-dimensional model and and now put 230 00:08:49,610 --> 00:08:45,000 it in the case of moist convection we 231 00:08:52,730 --> 00:08:49,620 get this this profile here and this as I 232 00:08:55,190 --> 00:08:52,740 mentioned is representing temperature 233 00:08:56,570 --> 00:08:55,200 that's decreasing not quite as strongly 234 00:08:59,090 --> 00:08:56,580 with height and this is because again 235 00:09:01,400 --> 00:08:59,100 the significant heating due to the space 236 00:09:03,200 --> 00:09:01,410 water reduces the temperature lapse 237 00:09:05,060 --> 00:09:03,210 right you don't cool as much as you're 238 00:09:07,040 --> 00:09:05,070 lifting up and so this allows the 239 00:09:09,080 --> 00:09:07,050 atmosphere to come back to a more stable 240 00:09:11,960 --> 00:09:09,090 lapse rate than if we didn't have 241 00:09:13,910 --> 00:09:11,970 moisture involved the fact that we have 242 00:09:15,620 --> 00:09:13,920 moist convection also cools the surface 243 00:09:18,320 --> 00:09:15,630 a bit so now we're at a realistic value 244 00:09:20,570 --> 00:09:18,330 of around 300 Kelvin for our average 245 00:09:23,360 --> 00:09:20,580 surface temperature so that's an 246 00:09:24,890 --> 00:09:23,370 important feature as well there are 247 00:09:26,750 --> 00:09:24,900 several other consequences of the fact 248 00:09:28,700 --> 00:09:26,760 that the conduction that happens is 249 00:09:29,120 --> 00:09:28,710 moist involves the phase changes in 250 00:09:34,400 --> 00:09:29,130 water 251 00:09:36,470 --> 00:09:34,410 which is an important greenhouse gas in 252 00:09:38,150 --> 00:09:36,480 the atmosphere so that has radiative 253 00:09:39,770 --> 00:09:38,160 consequences and it this was the only 254 00:09:41,630 --> 00:09:39,780 thing that moist convection did it would 255 00:09:43,400 --> 00:09:41,640 still be very important for us to 256 00:09:45,380 --> 00:09:43,410 consider but it's not the only thing it 257 00:09:47,960 --> 00:09:45,390 does it forms clouds which themselves 258 00:09:50,840 --> 00:09:47,970 affect the albedo and long-wave trapping 259 00:09:52,220 --> 00:09:50,850 and have huge radiative impacts another 260 00:09:54,230 --> 00:09:52,230 sort of unique feature of moist 261 00:09:56,570 --> 00:09:54,240 convection is that it causes us to have 262 00:09:58,550 --> 00:09:56,580 widely spaced updrafts if this is just a 263 00:10:01,040 --> 00:09:58,560 picture from an airplane of some 264 00:10:03,260 --> 00:10:01,050 tropical convection and you can see that 265 00:10:04,820 --> 00:10:03,270 you know we have rising motions in these 266 00:10:06,890 --> 00:10:04,830 clouds here and then it's sort of beat 267 00:10:09,440 --> 00:10:06,900 there's these large areas of clear air 268 00:10:12,350 --> 00:10:09,450 which is sinking in between and this is 269 00:10:14,570 --> 00:10:12,360 because the amount the air that is 270 00:10:17,240 --> 00:10:14,580 sinking in those clear set areas the 271 00:10:19,370 --> 00:10:17,250 rate at which it sinks it's controlled 272 00:10:22,550 --> 00:10:19,380 by how how strongly the atmosphere is 273 00:10:24,680 --> 00:10:22,560 cooling in those regions and that is not 274 00:10:26,990 --> 00:10:24,690 a very very fast rate so they sink it 275 00:10:29,180 --> 00:10:27,000 sinks pretty slowly whereas in clouds 276 00:10:31,250 --> 00:10:29,190 the air rises pretty quickly and so if 277 00:10:33,260 --> 00:10:31,260 we have air that goes up quickly but 278 00:10:35,660 --> 00:10:33,270 down slowly we need to have a fractional 279 00:10:37,880 --> 00:10:35,670 area difference in the coverage of those 280 00:10:39,290 --> 00:10:37,890 areas so that to have mass balance so 281 00:10:41,030 --> 00:10:39,300 the air that has to go up has to go up 282 00:10:43,040 --> 00:10:41,040 in pretty narrow regions and the air 283 00:10:45,950 --> 00:10:43,050 that's going down has to cover large 284 00:10:47,870 --> 00:10:45,960 areas to balance and so this is a really 285 00:10:49,910 --> 00:10:47,880 sort of fundamental property of moist 286 00:10:51,380 --> 00:10:49,920 convection that is not the case for dry 287 00:10:53,000 --> 00:10:51,390 convection and dry convection you can 288 00:10:56,840 --> 00:10:53,010 have symmetry between your up-and-down 289 00:10:59,270 --> 00:10:56,850 motions so keeping all this in mind then 290 00:11:00,800 --> 00:10:59,280 what are the properties of our state if 291 00:11:02,720 --> 00:11:00,810 we consider the atmosphere to be in this 292 00:11:05,720 --> 00:11:02,730 state of radiant convective equilibrium 293 00:11:07,520 --> 00:11:05,730 or RCE well again we have this balance 294 00:11:10,250 --> 00:11:07,530 between net rate of cooling and the 295 00:11:12,070 --> 00:11:10,260 turbulent fluxes the feet one important 296 00:11:13,900 --> 00:11:12,080 things to remember is that 297 00:11:16,840 --> 00:11:13,910 in this this way of thinking about it 298 00:11:19,030 --> 00:11:16,850 convection is a mixing process it's not 299 00:11:21,040 --> 00:11:19,040 some externalized heat source that just 300 00:11:23,440 --> 00:11:21,050 pumps heat into your system no it moves 301 00:11:25,750 --> 00:11:23,450 around heat in the in the vertical and 302 00:11:28,360 --> 00:11:25,760 so it's you cannot just assume it as as 303 00:11:31,570 --> 00:11:28,370 some sort of import of heat into your 304 00:11:33,430 --> 00:11:31,580 system and it arranges itself to balance 305 00:11:34,990 --> 00:11:33,440 the radiation so here the you know the 306 00:11:36,910 --> 00:11:35,000 reading of cooling is being driven by 307 00:11:38,650 --> 00:11:36,920 you know external things like the Sun 308 00:11:40,750 --> 00:11:38,660 and the constituents at the atmosphere 309 00:11:42,640 --> 00:11:40,760 in convection just does what it needs to 310 00:11:44,740 --> 00:11:42,650 in order to balance that radiative 311 00:11:46,390 --> 00:11:44,750 cooling you have this battle between 312 00:11:48,490 --> 00:11:46,400 convection which is trying to make 313 00:11:50,320 --> 00:11:48,500 things neutral to convection and 314 00:11:52,240 --> 00:11:50,330 radiation which is trying to push you 315 00:11:54,420 --> 00:11:52,250 towards a radiative equilibrium which is 316 00:11:56,920 --> 00:11:54,430 very unstable and in that battle 317 00:11:59,620 --> 00:11:56,930 convection wins because convection is a 318 00:12:01,780 --> 00:11:59,630 fast process whereas radiation is slow I 319 00:12:03,670 --> 00:12:01,790 mean the radiative relaxation timescale 320 00:12:05,830 --> 00:12:03,680 in Earth's atmosphere is on the order of 321 00:12:07,900 --> 00:12:05,840 a month or so where is the convective 322 00:12:10,570 --> 00:12:07,910 timescales like an hour okay so that's 323 00:12:12,940 --> 00:12:10,580 why our mean temperature profile is in 324 00:12:15,400 --> 00:12:12,950 this this balance state are the ones 325 00:12:17,680 --> 00:12:15,410 that are responding to a neutral state 326 00:12:20,590 --> 00:12:17,690 or convection and this has important 327 00:12:22,180 --> 00:12:20,600 constraints then on the basic properties 328 00:12:24,670 --> 00:12:22,190 of the climate system it contains our 329 00:12:26,740 --> 00:12:24,680 lapse rate that decrease in temperature 330 00:12:28,570 --> 00:12:26,750 with altitude to be moist adiabatic 331 00:12:30,940 --> 00:12:28,580 which has other consequences for how 332 00:12:32,830 --> 00:12:30,950 things depend on warming so for example 333 00:12:34,600 --> 00:12:32,840 that that rate changes with warming 334 00:12:36,880 --> 00:12:34,610 because again the moist convective lapse 335 00:12:37,960 --> 00:12:36,890 rate changes warming and it causes if 336 00:12:40,540 --> 00:12:37,970 we're thinking about future climate 337 00:12:42,580 --> 00:12:40,550 change for example it causes the upper 338 00:12:43,900 --> 00:12:42,590 troposphere the temperature that it you 339 00:12:46,030 --> 00:12:43,910 know fifteen twenty twelve in the 340 00:12:49,210 --> 00:12:46,040 tropics to warm more than they do near 341 00:12:51,160 --> 00:12:49,220 the surface now this again is a 342 00:12:53,260 --> 00:12:51,170 statistical equilibrium it holds on 343 00:12:54,580 --> 00:12:53,270 average over a large area and timescale 344 00:12:57,340 --> 00:12:54,590 it doesn't hold it any given particular 345 00:12:58,980 --> 00:12:57,350 location and it's upset by large-scale 346 00:13:01,330 --> 00:12:58,990 circulation so if you have lateral 347 00:13:04,210 --> 00:13:01,340 transport of energy in the horizontal 348 00:13:05,890 --> 00:13:04,220 then you don't necessarily have this as 349 00:13:08,280 --> 00:13:05,900 holding but if you integrate over those 350 00:13:10,780 --> 00:13:08,290 circulations it works on average in 351 00:13:13,420 --> 00:13:10,790 addition one really fundamental property 352 00:13:15,300 --> 00:13:13,430 is that in the system the radiation and 353 00:13:17,890 --> 00:13:15,310 the convection are highly interactive 354 00:13:20,410 --> 00:13:17,900 okay because radiation the rated cooling 355 00:13:23,040 --> 00:13:20,420 rates depend on water vapor and clouds 356 00:13:25,650 --> 00:13:23,050 and convection controls the disk 357 00:13:27,870 --> 00:13:25,660 Bhushan of water vapor in clouds and so 358 00:13:30,180 --> 00:13:27,880 this interact interaction between the 359 00:13:31,860 --> 00:13:30,190 two gives us the potential for 360 00:13:34,170 --> 00:13:31,870 instability and that's what I'm going to 361 00:13:38,550 --> 00:13:34,180 talk kind of in the rest of the talk 362 00:13:41,370 --> 00:13:38,560 about so we can simulate this sort of 363 00:13:43,019 --> 00:13:41,380 state in a variety of different modeling 364 00:13:45,990 --> 00:13:43,029 frameworks and there's been a long 365 00:13:46,740 --> 00:13:46,000 history of doing this starting with 366 00:13:50,420 --> 00:13:46,750 these sort of one-dimensional 367 00:13:53,130 --> 00:13:50,430 simulations I talked about in the 1960s 368 00:13:54,680 --> 00:13:53,140 but then moving on to two-dimensional 369 00:13:57,630 --> 00:13:54,690 simulations where you're actually 370 00:13:59,190 --> 00:13:57,640 explicitly resolving the clouds and 371 00:14:01,500 --> 00:13:59,200 convection and convective heat transport 372 00:14:03,329 --> 00:14:01,510 and then more recently in three 373 00:14:04,710 --> 00:14:03,339 dimensional simulations and so in a 374 00:14:06,019 --> 00:14:04,720 three dimensional simulation how do you 375 00:14:08,460 --> 00:14:06,029 simulate RCE 376 00:14:11,069 --> 00:14:08,470 well you consider usually some sort of 377 00:14:13,710 --> 00:14:11,079 doubly periodic box that's hundreds of 378 00:14:15,360 --> 00:14:13,720 kilometers in each dimension and you 379 00:14:18,509 --> 00:14:15,370 know covers the depth of the troposphere 380 00:14:20,490 --> 00:14:18,519 plus a bit more at the top of the model 381 00:14:21,720 --> 00:14:20,500 we have solar insulation coming in and 382 00:14:23,699 --> 00:14:21,730 we constrain that to be the same 383 00:14:26,069 --> 00:14:23,709 everywhere across the domains 384 00:14:28,019 --> 00:14:26,079 there's no gradients in in the solar 385 00:14:30,060 --> 00:14:28,029 insulation and the lower boundary 386 00:14:32,069 --> 00:14:30,070 condition is also homogeneous generally 387 00:14:33,930 --> 00:14:32,079 it's fixed sea surface temperature and 388 00:14:35,940 --> 00:14:33,940 we make that the same value everywhere 389 00:14:37,290 --> 00:14:35,950 so everything is the same everywhere 390 00:14:39,269 --> 00:14:37,300 across the domain in terms of the 391 00:14:41,730 --> 00:14:39,279 boundary conditions of forcing and we 392 00:14:43,230 --> 00:14:41,740 just initialize it with random noise in 393 00:14:45,449 --> 00:14:43,240 the boundary layer just to kind of get 394 00:14:48,060 --> 00:14:45,459 things moving and it begins to convict 395 00:14:49,889 --> 00:14:48,070 and overturn and then those clouds and 396 00:14:51,389 --> 00:14:49,899 convection are resolved explicitly by 397 00:14:53,490 --> 00:14:51,399 you know the equation governing 398 00:14:55,139 --> 00:14:53,500 equations in this model which are you 399 00:14:58,079 --> 00:14:55,149 know the basic fundamental dynamical 400 00:15:01,170 --> 00:14:58,089 equations of motion and we can see what 401 00:15:03,150 --> 00:15:01,180 state it ends up in so the best way to 402 00:15:05,880 --> 00:15:03,160 tell you to show you that is to actually 403 00:15:07,889 --> 00:15:05,890 just show you so this is a movie of one 404 00:15:09,689 --> 00:15:07,899 of these cloud resolving simulations of 405 00:15:11,880 --> 00:15:09,699 radiative equilibrium we're looking down 406 00:15:14,460 --> 00:15:11,890 here on the model it's about a thousand 407 00:15:17,220 --> 00:15:14,470 kilometers in each dimension the grey 408 00:15:20,100 --> 00:15:17,230 shading our clouds so it's basically one 409 00:15:22,380 --> 00:15:20,110 isosurface of cloud condensate the red 410 00:15:25,920 --> 00:15:22,390 colors the color shading is humidity 411 00:15:28,079 --> 00:15:25,930 near the surface where red is kind of 412 00:15:30,630 --> 00:15:28,089 moist air and blue as you can see 413 00:15:32,040 --> 00:15:30,640 developing here is dry so initially as 414 00:15:33,480 --> 00:15:32,050 you can see the clouds and convection 415 00:15:35,280 --> 00:15:33,490 we're kind of randomly distributed it 416 00:15:36,900 --> 00:15:35,290 looks like popcorn or fireworks or 417 00:15:37,410 --> 00:15:36,910 something like that but that doesn't 418 00:15:40,110 --> 00:15:37,420 stay 419 00:15:42,780 --> 00:15:40,120 that case and particularly see this area 420 00:15:44,910 --> 00:15:42,790 of dryer air that starts to form and we 421 00:15:47,220 --> 00:15:44,920 don't have clouds and convection there 422 00:15:50,070 --> 00:15:47,230 anymore because thunderstorms don't like 423 00:15:52,860 --> 00:15:50,080 it when there's dry air and this dry 424 00:15:54,900 --> 00:15:52,870 patch gets drier and drier with time and 425 00:15:56,400 --> 00:15:54,910 it and it starts to expand and so this 426 00:15:59,280 --> 00:15:56,410 is what is leading to the self 427 00:16:01,980 --> 00:15:59,290 aggregation process as that happens the 428 00:16:03,960 --> 00:16:01,990 moist areas of the domain get moister 429 00:16:05,550 --> 00:16:03,970 where's the where's the dry areas get 430 00:16:08,250 --> 00:16:05,560 drier so you have this net sort of 431 00:16:09,810 --> 00:16:08,260 transport of moist static energy and 432 00:16:12,570 --> 00:16:09,820 moisture from the dry regions to the 433 00:16:14,730 --> 00:16:12,580 moist regions and as the dry region 434 00:16:16,470 --> 00:16:14,740 expands convection is suppressed there 435 00:16:19,080 --> 00:16:16,480 and therefore becomes increasingly 436 00:16:21,600 --> 00:16:19,090 localized into a single or in other 437 00:16:23,520 --> 00:16:21,610 models perhaps multiple intensely 438 00:16:25,950 --> 00:16:23,530 precipitating clusters so all the clouds 439 00:16:28,050 --> 00:16:25,960 and precipitation and rising motion is 440 00:16:31,080 --> 00:16:28,060 confined in this one area and the rest 441 00:16:33,960 --> 00:16:31,090 of the domain has drier air that is 442 00:16:35,580 --> 00:16:33,970 clear and slowly sinking and so this is 443 00:16:38,160 --> 00:16:35,590 known as self aggregation it's this 444 00:16:40,320 --> 00:16:38,170 spontaneous transition from randomly 445 00:16:41,820 --> 00:16:40,330 distributed to organized convection so 446 00:16:44,580 --> 00:16:41,830 there's nothing that said oh I need to 447 00:16:46,230 --> 00:16:44,590 have my clouds here all the conditions 448 00:16:48,540 --> 00:16:46,240 were the same everywhere it decided to 449 00:16:50,760 --> 00:16:48,550 do that on its own and what causes it to 450 00:16:53,520 --> 00:16:50,770 do that well I'm gonna explain this in 451 00:16:56,340 --> 00:16:53,530 more detail but it results from 452 00:16:59,100 --> 00:16:56,350 interactions between the convection and 453 00:17:01,920 --> 00:16:59,110 its environment involving clouds water 454 00:17:04,680 --> 00:17:01,930 vapor radiation surface fluxes and the 455 00:17:06,870 --> 00:17:04,690 circulation oh this localization of 456 00:17:10,710 --> 00:17:06,880 connection was first seen in a paper by 457 00:17:12,090 --> 00:17:10,720 Isaac held in 1993 and I'm gonna attempt 458 00:17:14,640 --> 00:17:12,100 in the rest of my talk to summarize a 459 00:17:16,920 --> 00:17:14,650 couple recent reviews that I wrote on 460 00:17:18,300 --> 00:17:16,930 the topic okay so let's try to better 461 00:17:20,130 --> 00:17:18,310 understand what what this self 462 00:17:22,800 --> 00:17:20,140 aggregation phenomena is what causes it 463 00:17:24,689 --> 00:17:22,810 and what impact does it does it have me 464 00:17:26,630 --> 00:17:24,699 does it matter that the convection is 465 00:17:28,980 --> 00:17:26,640 organized in a clump like this I mean if 466 00:17:30,690 --> 00:17:28,990 you're in this spot and it's raining on 467 00:17:32,850 --> 00:17:30,700 you it matters but does it matter for 468 00:17:35,310 --> 00:17:32,860 the larger scale climate of the 469 00:17:37,500 --> 00:17:35,320 atmosphere and the first thing I like to 470 00:17:39,780 --> 00:17:37,510 say is that this is not a peculiarity of 471 00:17:41,850 --> 00:17:39,790 just this one particular model or this 472 00:17:44,550 --> 00:17:41,860 one particular simulation this occurs in 473 00:17:47,100 --> 00:17:44,560 a wide variety of models and model 474 00:17:48,900 --> 00:17:47,110 configuration so it was first seen in 475 00:17:50,730 --> 00:17:48,910 two dimensional cloud resolving models 476 00:17:52,500 --> 00:17:50,740 this held paper I mentioned 477 00:17:58,049 --> 00:17:52,510 so this is just showing an image from 478 00:18:00,659 --> 00:17:58,059 that paper the why this is the spatial 479 00:18:03,899 --> 00:18:00,669 axis of the of the model and then it's 480 00:18:07,200 --> 00:18:03,909 these are time going forward the shading 481 00:18:08,760 --> 00:18:07,210 here is precipitation rainfall when CDC 482 00:18:10,529 --> 00:18:08,770 initially it sort of scattered about but 483 00:18:12,360 --> 00:18:10,539 then it gets organized into these two 484 00:18:14,340 --> 00:18:12,370 kind of clumps that stay in the same 485 00:18:17,549 --> 00:18:14,350 place as time progresses and eventually 486 00:18:19,019 --> 00:18:17,559 collapses only to one that's two 487 00:18:20,399 --> 00:18:19,029 dimensions what about three well I just 488 00:18:22,139 --> 00:18:20,409 showed you a three-dimensional movie but 489 00:18:24,570 --> 00:18:22,149 just to show a couple other examples 490 00:18:26,490 --> 00:18:24,580 this has been found to occur and a lot 491 00:18:28,470 --> 00:18:26,500 of different small domain square 492 00:18:31,110 --> 00:18:28,480 three-dimensional cloud resolving models 493 00:18:33,000 --> 00:18:31,120 this was the first one Tompkins in Craig 494 00:18:35,730 --> 00:18:33,010 98 we're looking just the left panel 495 00:18:37,500 --> 00:18:35,740 here what's plotted is a measure of 496 00:18:41,850 --> 00:18:37,510 humidity and so you see these kind of 497 00:18:43,799 --> 00:18:41,860 alternating moist and dry regions this 498 00:18:47,010 --> 00:18:43,809 was a really seminal paper on the topic 499 00:18:50,220 --> 00:18:47,020 by Chris Brotherton at all in 2005 where 500 00:18:52,470 --> 00:18:50,230 again you have all of your precipitation 501 00:18:54,630 --> 00:18:52,480 confined in this one really strongly 502 00:18:56,130 --> 00:18:54,640 raining cluster that it's very moist 503 00:18:58,350 --> 00:18:56,140 where the rest of the air surrounding it 504 00:19:02,070 --> 00:18:58,360 is dry and clear and cooling very 505 00:19:04,289 --> 00:19:02,080 strongly just space we also have seen it 506 00:19:06,899 --> 00:19:04,299 to occur in three-dimensional models of 507 00:19:08,669 --> 00:19:06,909 other geometries including ones that are 508 00:19:10,649 --> 00:19:08,679 long channels kind of like a bowling 509 00:19:12,299 --> 00:19:10,659 alley so this is an example from one of 510 00:19:14,549 --> 00:19:12,309 my papers this domain is three 511 00:19:16,830 --> 00:19:14,559 dimensional but in the horizontal it's 512 00:19:19,169 --> 00:19:16,840 12,000 kilometers long so that's you 513 00:19:21,690 --> 00:19:19,179 know substantial part of the way around 514 00:19:24,779 --> 00:19:21,700 the world but only 200 kilometers wide 515 00:19:26,760 --> 00:19:24,789 and what's shown here on the top is 516 00:19:28,950 --> 00:19:26,770 cloud the precipitation in the bottom is 517 00:19:31,049 --> 00:19:28,960 moisture and so again you can see you 518 00:19:33,180 --> 00:19:31,059 have kind of an area of clouds and 519 00:19:34,799 --> 00:19:33,190 precipitation and that a dry area and 520 00:19:36,870 --> 00:19:34,809 then another clump and then a dry area 521 00:19:38,909 --> 00:19:36,880 and so on and so forth and again you 522 00:19:40,049 --> 00:19:38,919 know here the boundary conditions and 523 00:19:41,669 --> 00:19:40,059 everything are exactly the same 524 00:19:44,850 --> 00:19:41,679 everywhere yet it yet it does this on 525 00:19:47,789 --> 00:19:44,860 its own we have also found that self 526 00:19:49,769 --> 00:19:47,799 aggregation occurs in regional and 527 00:19:51,690 --> 00:19:49,779 global models that have parameterize 528 00:19:53,789 --> 00:19:51,700 connections of them are running at a 529 00:19:56,310 --> 00:19:53,799 coarser spatial resolution so they have 530 00:19:58,889 --> 00:19:56,320 to parameterize convective motions and 531 00:20:00,960 --> 00:19:58,899 condensation and cloud formation this is 532 00:20:02,290 --> 00:20:00,970 an example of a couple different GCMs 533 00:20:04,030 --> 00:20:02,300 and these are the same 534 00:20:05,950 --> 00:20:04,040 models that are used for climate 535 00:20:08,530 --> 00:20:05,960 projections you know cmip5 that sort of 536 00:20:11,890 --> 00:20:08,540 thing but they're run here in this state 537 00:20:13,990 --> 00:20:11,900 where there's no land there's no word 538 00:20:15,790 --> 00:20:14,000 eotl insulation gradients there's no sea 539 00:20:17,440 --> 00:20:15,800 surface temperature gradients yet they 540 00:20:19,510 --> 00:20:17,450 still have this clustering of the 541 00:20:22,810 --> 00:20:19,520 precipitation into clumps around the 542 00:20:24,880 --> 00:20:22,820 world and then more recently we've also 543 00:20:27,400 --> 00:20:24,890 started to been able on with computing 544 00:20:29,470 --> 00:20:27,410 advances to do full global simulations 545 00:20:31,090 --> 00:20:29,480 but with models that explicitly resolve 546 00:20:34,690 --> 00:20:31,100 convection and so this is an example 547 00:20:36,130 --> 00:20:34,700 from the Japanese Nick ham group of a 548 00:20:38,880 --> 00:20:36,140 global simulation I think this was 549 00:20:43,420 --> 00:20:38,890 around ten kilometers and again you see 550 00:20:44,860 --> 00:20:43,430 clumps of precipitation forming now all 551 00:20:47,410 --> 00:20:44,870 of these simulations that I talked about 552 00:20:50,110 --> 00:20:47,420 this was in a configuration with no 553 00:20:51,550 --> 00:20:50,120 rotation so we stopped the earth from 554 00:20:53,200 --> 00:20:51,560 spinning and we're just it's just 555 00:20:54,700 --> 00:20:53,210 sitting there and so that's sort of 556 00:20:55,990 --> 00:20:54,710 representative of what things would be 557 00:20:58,570 --> 00:20:56,000 like near the equator because the 558 00:21:01,060 --> 00:20:58,580 Coriolis force is zero the equator but 559 00:21:03,760 --> 00:21:01,070 if we turn rotation back on and do 560 00:21:06,550 --> 00:21:03,770 rotating RCE on for example an F plane 561 00:21:08,380 --> 00:21:06,560 then we get a hurricane to form so the 562 00:21:11,140 --> 00:21:08,390 connection still self aggregates it just 563 00:21:14,290 --> 00:21:11,150 does so into a spinning cluster instead 564 00:21:15,880 --> 00:21:14,300 of a singular one and if you give it a 565 00:21:18,520 --> 00:21:15,890 high rotation rate you just have a world 566 00:21:22,150 --> 00:21:18,530 that fills with tropical cyclones which 567 00:21:23,680 --> 00:21:22,160 is pretty cool to watch actually okay so 568 00:21:26,650 --> 00:21:23,690 that's sort of what self aggregation 569 00:21:29,560 --> 00:21:26,660 looks like but how can we describe it 570 00:21:31,510 --> 00:21:29,570 more quantitatively well as I was 571 00:21:33,970 --> 00:21:31,520 telling you when we watch that movie as 572 00:21:35,980 --> 00:21:33,980 self aggregation progresses we have this 573 00:21:38,370 --> 00:21:35,990 moistening of the moist regions and 574 00:21:41,500 --> 00:21:38,380 drying of the dry region so we have a 575 00:21:45,070 --> 00:21:41,510 broadening of our humidity distribution 576 00:21:47,320 --> 00:21:45,080 one way to measure that is by something 577 00:21:49,330 --> 00:21:47,330 called the moist static energy or frozen 578 00:21:52,360 --> 00:21:49,340 like static energy variance and this is 579 00:21:53,950 --> 00:21:52,370 just a particular variable that depends 580 00:21:55,720 --> 00:21:53,960 on moisture in the atmosphere but it's 581 00:21:57,820 --> 00:21:55,730 an energetic quantity and so we like to 582 00:21:59,980 --> 00:21:57,830 use it because it's something that's 583 00:22:01,690 --> 00:21:59,990 conserved in the atmosphere so we always 584 00:22:03,850 --> 00:22:01,700 like to have conserved quantities 585 00:22:06,100 --> 00:22:03,860 they're easier to track in particular 586 00:22:08,770 --> 00:22:06,110 this is the equation here so it depends 587 00:22:11,320 --> 00:22:08,780 on the sum of internal and potential 588 00:22:13,720 --> 00:22:11,330 energy and then latent energy here 589 00:22:15,340 --> 00:22:13,730 including some contributions from ice 590 00:22:17,950 --> 00:22:15,350 that's why it's called 591 00:22:19,779 --> 00:22:17,960 like static energy and the important 592 00:22:21,159 --> 00:22:19,789 thing is that if we take the column 593 00:22:23,230 --> 00:22:21,169 integral of this integrated over the 594 00:22:25,120 --> 00:22:23,240 whole depth of the atmosphere that 595 00:22:27,010 --> 00:22:25,130 column integral is not changed by 596 00:22:29,080 --> 00:22:27,020 convection connection just redistributes 597 00:22:31,210 --> 00:22:29,090 my static energy in the vertical so when 598 00:22:32,830 --> 00:22:31,220 we do a budget for this quantity it's 599 00:22:34,720 --> 00:22:32,840 easier because we don't have to deal 600 00:22:38,680 --> 00:22:34,730 with convective tendencies which can be 601 00:22:40,210 --> 00:22:38,690 difficult to measure now in the tropical 602 00:22:42,970 --> 00:22:40,220 atmosphere most of the spatial 603 00:22:45,669 --> 00:22:42,980 variability in this quantity comes from 604 00:22:47,230 --> 00:22:45,679 the moisture term so this reflects this 605 00:22:49,240 --> 00:22:47,240 moist regions getting moister dry 606 00:22:53,049 --> 00:22:49,250 regions getting drier that I talked 607 00:22:56,529 --> 00:22:53,059 about so this is the axes on my plot 608 00:22:58,090 --> 00:22:56,539 disappeared but this is time from the 609 00:23:00,520 --> 00:22:58,100 beginning of the simulation to the end 610 00:23:02,529 --> 00:23:00,530 and then the variance is increasing here 611 00:23:04,990 --> 00:23:02,539 it's on a log scale so each of these 612 00:23:07,299 --> 00:23:05,000 kind of larger ticks is an order of 613 00:23:09,159 --> 00:23:07,309 magnitude and these are all just a 614 00:23:10,810 --> 00:23:09,169 different a couple sets of simulations 615 00:23:12,640 --> 00:23:10,820 and so we can see that in all of them 616 00:23:15,370 --> 00:23:12,650 the moist attic energy variance 617 00:23:17,380 --> 00:23:15,380 increases rapidly in the simulation so 618 00:23:18,039 --> 00:23:17,390 this is reflecting these moist regions 619 00:23:19,779 --> 00:23:18,049 gettin moister 620 00:23:21,340 --> 00:23:19,789 dry regions getting drier and it 621 00:23:26,320 --> 00:23:21,350 increases about two orders of magnitude 622 00:23:28,930 --> 00:23:26,330 so this is a big change okay now since 623 00:23:31,690 --> 00:23:28,940 we have this quantity that increases as 624 00:23:34,270 --> 00:23:31,700 the convection aggregates that means 625 00:23:36,279 --> 00:23:34,280 that a process that is acting to 626 00:23:39,370 --> 00:23:36,289 increase that variance is a process that 627 00:23:40,930 --> 00:23:39,380 favors the self aggregation okay so this 628 00:23:43,149 --> 00:23:40,940 is kind of begging for us to actually do 629 00:23:46,000 --> 00:23:43,159 a budget for this quantity and that can 630 00:23:47,890 --> 00:23:46,010 tell us what what mechanisms are leading 631 00:23:49,810 --> 00:23:47,900 to the self aggregation in these cases 632 00:23:51,940 --> 00:23:49,820 and so if we do a budget for this 633 00:23:55,169 --> 00:23:51,950 spatial variance of this more static 634 00:23:57,250 --> 00:23:55,179 energy economy this is what we get and 635 00:23:59,500 --> 00:23:57,260 basically the important thing here is 636 00:24:01,029 --> 00:23:59,510 that each of the so here's the tendency 637 00:24:03,039 --> 00:24:01,039 of the variance on the left hand side 638 00:24:05,680 --> 00:24:03,049 and each of the terms on the right hand 639 00:24:08,789 --> 00:24:05,690 side has to do with a source or sink a 640 00:24:12,700 --> 00:24:08,799 voice netic energy so what are those 641 00:24:16,000 --> 00:24:12,710 well first we have our turbulent fluxes 642 00:24:18,370 --> 00:24:16,010 of moist enthalpy from the sea surface 643 00:24:21,370 --> 00:24:18,380 to the atmosphere this is things like 644 00:24:24,220 --> 00:24:21,380 evaporation and sensible heat flux we 645 00:24:25,899 --> 00:24:24,230 then have our short wave component so 646 00:24:28,060 --> 00:24:25,909 this is the column short a flux 647 00:24:28,440 --> 00:24:28,070 convergent so the difference between the 648 00:24:30,240 --> 00:24:28,450 shore 649 00:24:32,160 --> 00:24:30,250 creative fluxes at the top of atmosphere 650 00:24:33,780 --> 00:24:32,170 and the surface then we have the same 651 00:24:35,010 --> 00:24:33,790 for the long ways again difference 652 00:24:37,560 --> 00:24:35,020 between talked about my sphere in 653 00:24:39,480 --> 00:24:37,570 surface and then we have the term having 654 00:24:41,670 --> 00:24:39,490 to do with advection so this is lateral 655 00:24:44,430 --> 00:24:41,680 transport of my psychic energy in or out 656 00:24:46,980 --> 00:24:44,440 of our column and so each of the terms 657 00:24:49,620 --> 00:24:46,990 in this budget is a feedback term where 658 00:24:51,900 --> 00:24:49,630 we have the product of an anomaly in 659 00:24:54,570 --> 00:24:51,910 moist attic energy itself this H prime 660 00:24:59,010 --> 00:24:54,580 and an anomaly in one of these sources 661 00:25:01,020 --> 00:24:59,020 or sinks so if we have a process that is 662 00:25:03,480 --> 00:25:01,030 acting to increase the moist attic 663 00:25:05,160 --> 00:25:03,490 energy of an already moist region but 664 00:25:07,200 --> 00:25:05,170 they we have more evaporation where 665 00:25:09,330 --> 00:25:07,210 where we already have lots of clouds and 666 00:25:11,640 --> 00:25:09,340 moisture and that's going to act to 667 00:25:13,980 --> 00:25:11,650 increase that moisture anomaly and cause 668 00:25:15,240 --> 00:25:13,990 a positive tendency in its variance and 669 00:25:17,190 --> 00:25:15,250 so therefore we call that a positive 670 00:25:19,860 --> 00:25:17,200 feedback and say that it favors 671 00:25:22,290 --> 00:25:19,870 self aggregation however if we have a 672 00:25:24,180 --> 00:25:22,300 process that is acting to decrease the 673 00:25:25,980 --> 00:25:24,190 moist attic energy of a moist region 674 00:25:29,670 --> 00:25:25,990 then that would damp that anomaly and we 675 00:25:32,520 --> 00:25:29,680 say that that would act against self 676 00:25:34,890 --> 00:25:32,530 aggregation as a negative feedback so we 677 00:25:36,960 --> 00:25:34,900 can calculate this budget in our model 678 00:25:39,420 --> 00:25:36,970 simulations and integrate it over the 679 00:25:41,250 --> 00:25:39,430 whole domain and say okay what processes 680 00:25:42,510 --> 00:25:41,260 are contributing to this increase in 681 00:25:44,490 --> 00:25:42,520 moist attic energy variance what 682 00:25:47,340 --> 00:25:44,500 processes are leading to self 683 00:25:51,390 --> 00:25:47,350 aggregation and so we can do that here 684 00:25:54,240 --> 00:25:51,400 and again the x axis here is time in our 685 00:25:56,010 --> 00:25:54,250 simulation above the dashed line is 686 00:25:57,690 --> 00:25:56,020 positive below it's negative and just 687 00:25:58,890 --> 00:25:57,700 focus kind of at the beginning of the 688 00:26:01,080 --> 00:25:58,900 simulation cover interested in this 689 00:26:03,240 --> 00:26:01,090 initial sort of instigation of 690 00:26:04,680 --> 00:26:03,250 aggregation and so each of the liens is 691 00:26:07,080 --> 00:26:04,690 one of the terms in our budget and 692 00:26:09,210 --> 00:26:07,090 really the key thing to take away is 693 00:26:10,410 --> 00:26:09,220 that at the beginning of the process out 694 00:26:12,930 --> 00:26:10,420 of the colored lines 695 00:26:15,150 --> 00:26:12,940 it's the blue and the green one that are 696 00:26:17,370 --> 00:26:15,160 positive and strongest and so that's the 697 00:26:18,540 --> 00:26:17,380 long wave and surface selects feedbacks 698 00:26:21,030 --> 00:26:18,550 and so we say that those are the 699 00:26:23,730 --> 00:26:21,040 processes that are driving the initial 700 00:26:25,260 --> 00:26:23,740 development of aggregation so what does 701 00:26:28,110 --> 00:26:25,270 that mean what physically are these 702 00:26:30,390 --> 00:26:28,120 feedbacks representing how can we have 703 00:26:32,640 --> 00:26:30,400 them be positive so first let me talk 704 00:26:34,260 --> 00:26:32,650 about the long wave feedback how do we 705 00:26:36,090 --> 00:26:34,270 get a positive long light radiation 706 00:26:39,080 --> 00:26:36,100 feedback in this context remember this 707 00:26:41,430 --> 00:26:39,090 is the feedback on these sort of spatial 708 00:26:44,999 --> 00:26:41,440 gradients a feedback on 709 00:26:47,460 --> 00:26:45,009 segregation well if we think about some 710 00:26:49,649 --> 00:26:47,470 parts of the atmosphere that are bit 711 00:26:52,499 --> 00:26:49,659 moisture than other parts that are dry 712 00:26:54,090 --> 00:26:52,509 we have preferentially our clouds and 713 00:26:56,580 --> 00:26:54,100 convection occurring in the moist 714 00:26:58,919 --> 00:26:56,590 regions so we have more clouds or water 715 00:27:01,860 --> 00:26:58,929 vapor in the moist regions than we do in 716 00:27:04,139 --> 00:27:01,870 the dry regions and so that means that 717 00:27:05,460 --> 00:27:04,149 the atmosphere in the dry regions you 718 00:27:07,200 --> 00:27:05,470 know there might be some shallow clouds 719 00:27:09,840 --> 00:27:07,210 or something the atmosphere there is 720 00:27:11,159 --> 00:27:09,850 cooling very strongly okay because a lot 721 00:27:13,409 --> 00:27:11,169 of the radiation emitted from the 722 00:27:15,480 --> 00:27:13,419 surface and these low clouds can readily 723 00:27:18,990 --> 00:27:15,490 escape to space through because it's 724 00:27:21,570 --> 00:27:19,000 very relatively transparent to radiation 725 00:27:24,749 --> 00:27:21,580 when it's very dry so dry regions are 726 00:27:26,789 --> 00:27:24,759 cooling very strongly whereas in the 727 00:27:28,409 --> 00:27:26,799 moist regions we have trapping of that 728 00:27:30,539 --> 00:27:28,419 long link radiation by you know the 729 00:27:32,399 --> 00:27:30,549 local greenhouse effect from all of the 730 00:27:34,379 --> 00:27:32,409 clouds and humidity there and so the 731 00:27:37,019 --> 00:27:34,389 atmosphere itself in these moist regions 732 00:27:38,460 --> 00:27:37,029 is cooling much less so we have small 733 00:27:41,999 --> 00:27:38,470 cooling in the moist regions large 734 00:27:44,340 --> 00:27:42,009 cooling in the dry regions and this wick 735 00:27:46,259 --> 00:27:44,350 will act to amplify those existing 736 00:27:48,450 --> 00:27:46,269 anomalies can where we have large 737 00:27:50,970 --> 00:27:48,460 amounts of cooling we're going to cause 738 00:27:52,529 --> 00:27:50,980 more subsidence which is going to dry 739 00:27:55,830 --> 00:27:52,539 the atmosphere even more make it more 740 00:27:57,749 --> 00:27:55,840 unfavorable to convection and cause a 741 00:27:59,759 --> 00:27:57,759 further decrease in my static energy 742 00:28:01,289 --> 00:27:59,769 whereas in these moist regions where 743 00:28:02,639 --> 00:28:01,299 it's not cooling very much that means 744 00:28:04,769 --> 00:28:02,649 it's going to be favorable for 745 00:28:06,629 --> 00:28:04,779 convection to continue to occur there 746 00:28:08,820 --> 00:28:06,639 and so this is a positive feedback 747 00:28:11,399 --> 00:28:08,830 because we're amplifying the existing 748 00:28:13,409 --> 00:28:11,409 pattern and it has to do with the fact 749 00:28:15,509 --> 00:28:13,419 that convection is controlling the 750 00:28:17,369 --> 00:28:15,519 distribution of clouds and moisture but 751 00:28:19,080 --> 00:28:17,379 then that affects the radiative 752 00:28:22,529 --> 00:28:19,090 tendencies in a way that positively 753 00:28:24,389 --> 00:28:22,539 feeds back on the connection now I'll 754 00:28:27,240 --> 00:28:24,399 just say that this doesn't always have 755 00:28:30,060 --> 00:28:27,250 to be a positive feedback whether or not 756 00:28:32,100 --> 00:28:30,070 it actually is depends on the vertical 757 00:28:33,930 --> 00:28:32,110 structure of your humidity perturbations 758 00:28:36,119 --> 00:28:33,940 and you know what your kind of basic 759 00:28:39,690 --> 00:28:36,129 state was but when it turns out to be 760 00:28:41,999 --> 00:28:39,700 positive this is why so that's a long 761 00:28:43,769 --> 00:28:42,009 wave radiation feedback I also mentioned 762 00:28:45,659 --> 00:28:43,779 that the surface flux feedback was 763 00:28:47,220 --> 00:28:45,669 contributing to self aggregation so what 764 00:28:49,560 --> 00:28:47,230 does that mean what is a positive 765 00:28:51,629 --> 00:28:49,570 surface flux you back well here we're 766 00:28:53,220 --> 00:28:51,639 talking about surface enthalpy fluxes so 767 00:28:54,000 --> 00:28:53,230 these are again these turbulent fluxes 768 00:28:56,610 --> 00:28:54,010 of latent 769 00:28:59,220 --> 00:28:56,620 sensible heat from the sea surface to 770 00:29:00,600 --> 00:28:59,230 the atmosphere so if we get again 771 00:29:02,460 --> 00:29:00,610 consider a little schematic where we 772 00:29:05,490 --> 00:29:02,470 have more clouds and convection in the 773 00:29:07,470 --> 00:29:05,500 moist areas that in the drive when we 774 00:29:09,600 --> 00:29:07,480 have more convection happening 775 00:29:11,490 --> 00:29:09,610 we have stronger these mean stronger 776 00:29:13,530 --> 00:29:11,500 winds at the surface of the atmosphere 777 00:29:14,520 --> 00:29:13,540 because you know you know that you've 778 00:29:16,080 --> 00:29:14,530 experienced this when a thunderstorm 779 00:29:18,690 --> 00:29:16,090 rolls through you have these strong 780 00:29:21,750 --> 00:29:18,700 gusts of wind and it's very turbulent 781 00:29:22,770 --> 00:29:21,760 for a while where is in the dry regions 782 00:29:24,450 --> 00:29:22,780 where we don't have a lot of convection 783 00:29:27,900 --> 00:29:24,460 going on it's pretty quiet and the winds 784 00:29:30,210 --> 00:29:27,910 are not very strong so more more winds 785 00:29:33,270 --> 00:29:30,220 more gust enos in the moist convection 786 00:29:35,340 --> 00:29:33,280 regions than in the dry and it turns out 787 00:29:37,890 --> 00:29:35,350 that enthalpy fluxes at the surface 788 00:29:39,480 --> 00:29:37,900 depend on wind speed okay this is called 789 00:29:41,730 --> 00:29:39,490 wishy the wind induced surface heat 790 00:29:44,130 --> 00:29:41,740 exchange effect and you know you know 791 00:29:46,470 --> 00:29:44,140 this if like this morning when I after I 792 00:29:48,539 --> 00:29:46,480 showered my hair was wet and I walked 793 00:29:50,370 --> 00:29:48,549 outside and as the air blew on as I go 794 00:29:52,680 --> 00:29:50,380 gosh I feel very cold because all this 795 00:29:54,210 --> 00:29:52,690 heat is evaporating from my head as the 796 00:29:54,659 --> 00:29:54,220 wind blow so that's why we use hair 797 00:29:58,200 --> 00:29:54,669 dryers 798 00:30:00,510 --> 00:29:58,210 you know transfers a evaporation of the 799 00:30:02,340 --> 00:30:00,520 water on our hair more efficiently so 800 00:30:04,470 --> 00:30:02,350 where we have stronger winds we have 801 00:30:06,930 --> 00:30:04,480 more evaporation stronger flux of heat 802 00:30:09,419 --> 00:30:06,940 from the sea surface to the atmosphere 803 00:30:11,549 --> 00:30:09,429 and so that large surface fluxes where 804 00:30:13,230 --> 00:30:11,559 its moist small words dry now there's 805 00:30:17,070 --> 00:30:13,240 other things that can affect this these 806 00:30:18,419 --> 00:30:17,080 fluxes also depend on the the enthalpy 807 00:30:20,039 --> 00:30:18,429 disequilibrium the difference in 808 00:30:22,770 --> 00:30:20,049 temperatures and humidity between the 809 00:30:25,320 --> 00:30:22,780 surface and atmosphere but at least in 810 00:30:28,049 --> 00:30:25,330 the beginning this wind effect winds 811 00:30:30,150 --> 00:30:28,059 basically and so we have more 812 00:30:31,740 --> 00:30:30,160 evaporation in the areas where it's 813 00:30:33,750 --> 00:30:31,750 already moist and so we're gonna moisten 814 00:30:37,830 --> 00:30:33,760 those so this is again a positive 815 00:30:39,930 --> 00:30:37,840 feedback on aggregation now these are 816 00:30:41,760 --> 00:30:39,940 not the only two processes that can 817 00:30:42,840 --> 00:30:41,770 contribute to aggregation they're kind 818 00:30:44,340 --> 00:30:42,850 of the most important ones but there are 819 00:30:46,530 --> 00:30:44,350 a couple other things that can happen 820 00:30:48,780 --> 00:30:46,540 there just to kind of summarize all of 821 00:30:52,320 --> 00:30:48,790 the processes that favor this cells 822 00:30:54,299 --> 00:30:52,330 aggregation clumping of convection well 823 00:30:56,159 --> 00:30:54,309 again we have our radiative feedbacks um 824 00:30:58,500 --> 00:30:56,169 the long way feedbacks which have 825 00:31:00,600 --> 00:30:58,510 contributions from just water vapor as 826 00:31:02,669 --> 00:31:00,610 well as clouds so we've got male water 827 00:31:04,820 --> 00:31:02,679 river and long wave clouds feedbacks 828 00:31:07,740 --> 00:31:04,830 which work pretty much the same way we 829 00:31:09,950 --> 00:31:07,750 have our surface flux feedback but again 830 00:31:14,039 --> 00:31:09,960 different by this wind speed variability 831 00:31:15,720 --> 00:31:14,049 but these radiative flux is have more 832 00:31:17,370 --> 00:31:15,730 than just a direct impact on the cooling 833 00:31:19,200 --> 00:31:17,380 in their column whenever you have 834 00:31:21,080 --> 00:31:19,210 spatial differences in the amount of 835 00:31:23,850 --> 00:31:21,090 graded cooling that can also drive 836 00:31:26,549 --> 00:31:23,860 lateral circulations okay at heating 837 00:31:29,210 --> 00:31:26,559 difference drives motion so we can also 838 00:31:31,289 --> 00:31:29,220 have a coupling between cooling 839 00:31:33,930 --> 00:31:31,299 particularly the strong cooling at the 840 00:31:35,580 --> 00:31:33,940 top of low clouds and a circulation and 841 00:31:37,620 --> 00:31:35,590 I'll explain more in a moment what that 842 00:31:40,500 --> 00:31:37,630 means and then we can also have 843 00:31:42,180 --> 00:31:40,510 contributions from the feedback between 844 00:31:43,649 --> 00:31:42,190 moisture and conductance I was sort of 845 00:31:45,060 --> 00:31:43,659 alluding to this and I'm saying well you 846 00:31:46,890 --> 00:31:45,070 know convection likes to occur where 847 00:31:48,510 --> 00:31:46,900 it's moist and where you have convection 848 00:31:50,789 --> 00:31:48,520 it moistens the atmosphere in more and 849 00:31:53,340 --> 00:31:50,799 so kind of convection who gets more 850 00:31:55,740 --> 00:31:53,350 convection in that sense but just to 851 00:31:57,810 --> 00:31:55,750 kind of explain this this radiative 852 00:32:00,360 --> 00:31:57,820 cooling circulation coupling a little 853 00:32:03,659 --> 00:32:00,370 bit more here's a kind of more detailed 854 00:32:06,210 --> 00:32:03,669 schematic of it the idea is again you 855 00:32:09,210 --> 00:32:06,220 have your deep convection and clouds in 856 00:32:10,850 --> 00:32:09,220 one area and then this drier area where 857 00:32:15,180 --> 00:32:10,860 maybe you just have some shallow clouds 858 00:32:17,250 --> 00:32:15,190 outside and basically in the dry region 859 00:32:19,440 --> 00:32:17,260 we have this large-scale sinking air 860 00:32:21,960 --> 00:32:19,450 large-scale subsidence in the atmosphere 861 00:32:24,960 --> 00:32:21,970 that promotes the formation of these low 862 00:32:28,220 --> 00:32:24,970 clouds and those low clouds cools very 863 00:32:30,060 --> 00:32:28,230 strongly at their tops okay so that's 864 00:32:34,680 --> 00:32:30,070 represented here blues little puffs of 865 00:32:37,080 --> 00:32:34,690 scribbles this low-level cooling drives 866 00:32:38,940 --> 00:32:37,090 a shallow circulation that increases the 867 00:32:41,220 --> 00:32:38,950 subsidence in these areas right because 868 00:32:42,899 --> 00:32:41,230 if it's if our apt if our air is cooling 869 00:32:45,330 --> 00:32:42,909 it needs to sink in order to balance 870 00:32:47,909 --> 00:32:45,340 that temperature change and so as it 871 00:32:50,430 --> 00:32:47,919 sinks then it has to spread out and then 872 00:32:51,930 --> 00:32:50,440 will rise back up in the convecting 873 00:32:54,110 --> 00:32:51,940 areas and so these are called 874 00:32:56,610 --> 00:32:54,120 radiatively driven cold fuels and 875 00:32:58,289 --> 00:32:56,620 because of the spatial differences in 876 00:33:01,049 --> 00:32:58,299 the cooling we get these circulations 877 00:33:03,570 --> 00:33:01,059 that form that will act to kind of give 878 00:33:05,039 --> 00:33:03,580 convection and an extra boost in those 879 00:33:07,529 --> 00:33:05,049 areas where it's already occurring and 880 00:33:10,799 --> 00:33:07,539 continue to damp it where it's not and 881 00:33:12,990 --> 00:33:10,809 so these these effects of the connection 882 00:33:15,450 --> 00:33:13,000 on the radiation have both direct effect 883 00:33:16,830 --> 00:33:15,460 effects locally and indirect effects 884 00:33:19,940 --> 00:33:16,840 because of how they can drive 885 00:33:24,950 --> 00:33:22,760 so that's why this convection can 886 00:33:27,680 --> 00:33:24,960 organize again it all depends on how the 887 00:33:29,840 --> 00:33:27,690 clouds and water vapor that are affected 888 00:33:32,000 --> 00:33:29,850 by the distribution section how they 889 00:33:34,550 --> 00:33:32,010 interact with circulations and these 890 00:33:36,200 --> 00:33:34,560 radiuses and so it's this because it's 891 00:33:37,520 --> 00:33:36,210 the interaction with convection in its 892 00:33:42,200 --> 00:33:37,530 own environment that's why it's this 893 00:33:44,030 --> 00:33:42,210 self aggregation we call so earlier I 894 00:33:45,950 --> 00:33:44,040 said and we want to know does this 895 00:33:48,020 --> 00:33:45,960 matter like why we care that comes 896 00:33:50,780 --> 00:33:48,030 clouds are clumped together well it 897 00:33:52,460 --> 00:33:50,790 turns out that self aggregation is not - 898 00:33:55,160 --> 00:33:52,470 just a spatial reorganization of 899 00:33:57,760 --> 00:33:55,170 connection but it has a large impact on 900 00:34:01,760 --> 00:33:57,770 the domain mean or large-scale climate 901 00:34:03,680 --> 00:34:01,770 okay so it has a kind of you know more 902 00:34:05,690 --> 00:34:03,690 fundamental effect on thing so what is 903 00:34:08,000 --> 00:34:05,700 that effect you know what what is 904 00:34:09,740 --> 00:34:08,010 different about the average state of 905 00:34:11,690 --> 00:34:09,750 that simulation when the convection is 906 00:34:13,820 --> 00:34:11,700 clumped together versus distributed 907 00:34:15,830 --> 00:34:13,830 randomly well the first one we've 908 00:34:17,620 --> 00:34:15,840 already talked about dry regions yet 909 00:34:20,270 --> 00:34:17,630 drier and moist regions get moister 910 00:34:21,950 --> 00:34:20,280 another way to view this is by looking 911 00:34:23,870 --> 00:34:21,960 at the spatial variance of humidity 912 00:34:26,240 --> 00:34:23,880 itself so this is column relative 913 00:34:29,510 --> 00:34:26,250 humidity which again in these 914 00:34:32,060 --> 00:34:29,520 simulations that start out from a random 915 00:34:34,130 --> 00:34:32,070 state as they organize in the convection 916 00:34:36,169 --> 00:34:34,140 aggregates we see a big increase in the 917 00:34:37,370 --> 00:34:36,179 spatial variance that occurs you know 918 00:34:39,080 --> 00:34:37,380 these are all simulations of different 919 00:34:41,870 --> 00:34:39,090 temperatures and it occurs in all of 920 00:34:43,490 --> 00:34:41,880 them okay so dry regions get dry or 921 00:34:44,360 --> 00:34:43,500 moist regions get more as the driver you 922 00:34:46,159 --> 00:34:44,370 can see a driver that means you're 923 00:34:48,169 --> 00:34:46,169 opening up more of these clear sky 924 00:34:51,980 --> 00:34:48,179 radiations that can cool more strongly 925 00:34:53,540 --> 00:34:51,990 in space now this is showing a column 926 00:34:56,000 --> 00:34:53,550 integrated measure but if we look at 927 00:34:58,970 --> 00:34:56,010 profiles of humidity it turns out that 928 00:35:01,550 --> 00:34:58,980 dry regions get drier at all levels okay 929 00:35:04,730 --> 00:35:01,560 and there's a lot of lines on this plot 930 00:35:07,880 --> 00:35:04,740 but the key thing is that the blue lines 931 00:35:10,700 --> 00:35:07,890 are measuring the profiles of humidity 932 00:35:12,200 --> 00:35:10,710 this is specific humidity at the 933 00:35:14,300 --> 00:35:12,210 function of pressure so the surface is 934 00:35:16,010 --> 00:35:14,310 down here the blue ones are the profiles 935 00:35:18,410 --> 00:35:16,020 immunity in the moist regions and the 936 00:35:21,800 --> 00:35:18,420 red are in the dry region and as we go 937 00:35:25,340 --> 00:35:21,810 from the pronounced down colors that's 938 00:35:27,290 --> 00:35:25,350 with time simulation so as we go along 939 00:35:29,690 --> 00:35:27,300 with truck time the moist region get 940 00:35:31,880 --> 00:35:29,700 moister and the dry regions get drier 941 00:35:33,590 --> 00:35:31,890 and as you can see that occurs at pretty 942 00:35:35,000 --> 00:35:33,600 much every every level 943 00:35:37,580 --> 00:35:35,010 including you know pretty close to the 944 00:35:39,530 --> 00:35:37,590 surface the strongest change though is 945 00:35:40,820 --> 00:35:39,540 in kind of the lower to mid troposphere 946 00:35:43,630 --> 00:35:40,830 if you look at this drawing I mean this 947 00:35:46,940 --> 00:35:43,640 is a big difference almost in some cases 948 00:35:48,350 --> 00:35:46,950 a factor of two or more this one 949 00:35:49,970 --> 00:35:48,360 simulation and the dotted line is a 950 00:35:52,310 --> 00:35:49,980 particularly extreme one where you lose 951 00:35:53,900 --> 00:35:52,320 almost all of the water vapor in the in 952 00:35:57,500 --> 00:35:53,910 these dry regions and so that you know 953 00:36:00,020 --> 00:35:57,510 has a really big impact okay so so this 954 00:36:02,510 --> 00:36:00,030 is reflecting this broadening of our 955 00:36:04,880 --> 00:36:02,520 distribution of humidity more moist and 956 00:36:07,490 --> 00:36:04,890 dryness but what about the mean state at 957 00:36:10,190 --> 00:36:07,500 all is the average dryer or moister than 958 00:36:11,690 --> 00:36:10,200 it was initially well you can kind of 959 00:36:13,550 --> 00:36:11,700 figure out the answer if I said that dry 960 00:36:16,220 --> 00:36:13,560 regions are getting drier and we saw 961 00:36:19,040 --> 00:36:16,230 from those movies that the dry areas 962 00:36:20,720 --> 00:36:19,050 cover a larger fraction of the domain it 963 00:36:23,600 --> 00:36:20,730 turns out that the mean state is drier 964 00:36:25,670 --> 00:36:23,610 and this we see in these simulations of 965 00:36:27,710 --> 00:36:25,680 self aggregation as well as in 966 00:36:29,720 --> 00:36:27,720 observations of aggregated convection 967 00:36:31,010 --> 00:36:29,730 and so in observations we can't say what 968 00:36:33,170 --> 00:36:31,020 specifically what's causing the 969 00:36:35,320 --> 00:36:33,180 convection to organize but if it is 970 00:36:38,570 --> 00:36:35,330 organized we also note that the average 971 00:36:40,640 --> 00:36:38,580 atmosphere around it is drier so here 972 00:36:43,100 --> 00:36:40,650 the black line is the simulation the red 973 00:36:44,870 --> 00:36:43,110 is observations dashed is when the 974 00:36:46,520 --> 00:36:44,880 convection is scattered and then the 975 00:36:48,380 --> 00:36:46,530 solid line is when it's organized and 976 00:36:51,860 --> 00:36:48,390 here it's in this case relative humidity 977 00:36:54,380 --> 00:36:51,870 that's plotted again in the lower to mid 978 00:36:56,990 --> 00:36:54,390 troposphere we see this big amount of 979 00:36:58,670 --> 00:36:57,000 drying as we go towards more aggregated 980 00:37:01,370 --> 00:36:58,680 conditions it's less extreme and 981 00:37:06,470 --> 00:37:01,380 observations but we still see a signal 982 00:37:08,270 --> 00:37:06,480 of the same sign so we thought drying 983 00:37:11,240 --> 00:37:08,280 the dry regions drawing up the mean 984 00:37:11,600 --> 00:37:11,250 state that's shown here on the right 985 00:37:13,790 --> 00:37:11,610 again 986 00:37:16,190 --> 00:37:13,800 but what about temperature changes well 987 00:37:18,350 --> 00:37:16,200 it seems that we also see a warming of 988 00:37:20,540 --> 00:37:18,360 the atmosphere when convection Agria so 989 00:37:22,640 --> 00:37:20,550 here these plots are showing differences 990 00:37:24,080 --> 00:37:22,650 between the end of the simulation when 991 00:37:26,390 --> 00:37:24,090 it's organized in the beginning when 992 00:37:28,520 --> 00:37:26,400 it's not and so if it's positive that 993 00:37:31,220 --> 00:37:28,530 means that when it is organized it is 994 00:37:34,160 --> 00:37:31,230 warmer here the sign is positive on the 995 00:37:37,190 --> 00:37:34,170 right and then again drier when we are 996 00:37:39,380 --> 00:37:37,200 aggregated and again my axis was cut off 997 00:37:41,830 --> 00:37:39,390 but this is a difference of a couple 998 00:37:44,210 --> 00:37:41,840 degrees Kelvin so that's significant 999 00:37:45,770 --> 00:37:44,220 actually it has to do from the fact that 1000 00:37:46,670 --> 00:37:45,780 when convection is clustered together 1001 00:37:53,480 --> 00:37:46,680 you 1002 00:37:56,930 --> 00:37:53,490 regions and so they're less susceptible 1003 00:37:58,730 --> 00:37:56,940 to entrainment because so you have less 1004 00:38:00,830 --> 00:37:58,740 mixing and they lose less buoyancy as 1005 00:38:03,350 --> 00:38:00,840 they rise and so you have you're able to 1006 00:38:05,240 --> 00:38:03,360 give back to a warmer temperature so 1007 00:38:08,810 --> 00:38:05,250 we've got drawing in the dry regions 1008 00:38:10,760 --> 00:38:08,820 drying of the mean state and warming all 1009 00:38:12,260 --> 00:38:10,770 right so that's already some substantial 1010 00:38:14,990 --> 00:38:12,270 changes any is there anything else going 1011 00:38:16,880 --> 00:38:15,000 on well there's also a change in the 1012 00:38:20,030 --> 00:38:16,890 distribution of clouds and in particular 1013 00:38:22,430 --> 00:38:20,040 on average there's a decrease in the 1014 00:38:23,840 --> 00:38:22,440 amount of high clouds with aggregation 1015 00:38:26,120 --> 00:38:23,850 and I show this a couple different ways 1016 00:38:27,890 --> 00:38:26,130 here these are again differences between 1017 00:38:29,420 --> 00:38:27,900 the end of the simulation when things 1018 00:38:32,780 --> 00:38:29,430 are clumped together and the beginning 1019 00:38:35,150 --> 00:38:32,790 when they are more scattered and on the 1020 00:38:37,070 --> 00:38:35,160 left is left panel is showing the actual 1021 00:38:40,430 --> 00:38:37,080 cloud fraction so the fraction of the 1022 00:38:43,100 --> 00:38:40,440 domain at any given level Heights that 1023 00:38:44,960 --> 00:38:43,110 is that it's covered by cloud and we've 1024 00:38:46,310 --> 00:38:44,970 got eight different simulations on here 1025 00:38:49,070 --> 00:38:46,320 so it's a little bit messy but you can 1026 00:38:50,930 --> 00:38:49,080 kind of see that you have this it's 1027 00:38:52,190 --> 00:38:50,940 negative basically in all of these who 1028 00:38:53,480 --> 00:38:52,200 have a decrease in the amount of high 1029 00:38:55,130 --> 00:38:53,490 cloud which you can see maybe a little 1030 00:38:57,110 --> 00:38:55,140 bit better in this pen on the right 1031 00:38:59,320 --> 00:38:57,120 which is showing the actual amount of 1032 00:39:02,180 --> 00:38:59,330 cloud water this is ice cloud 1033 00:39:05,450 --> 00:39:02,190 solid-phase condensate here which is all 1034 00:39:07,790 --> 00:39:05,460 decreasing at these high levels in terms 1035 00:39:10,850 --> 00:39:07,800 of low level clouds we see some 1036 00:39:13,040 --> 00:39:10,860 increases in the amount of low cloud 1037 00:39:14,330 --> 00:39:13,050 with aggregation but then there's some 1038 00:39:18,380 --> 00:39:14,340 simulations that seem to have a decrease 1039 00:39:20,330 --> 00:39:18,390 it's a little bit less clear and shallow 1040 00:39:21,650 --> 00:39:20,340 clouds in the atmosphere occur in kind 1041 00:39:24,260 --> 00:39:21,660 of small scales and so you need a really 1042 00:39:26,600 --> 00:39:24,270 high resolution to do them properly in 1043 00:39:29,030 --> 00:39:26,610 models and these simulations don't have 1044 00:39:30,590 --> 00:39:29,040 super high resolution it's about three 1045 00:39:32,480 --> 00:39:30,600 kilometers so I wouldn't really trust 1046 00:39:34,280 --> 00:39:32,490 the low clouds but the deep clouds are 1047 00:39:36,710 --> 00:39:34,290 pretty good and so we see this pretty 1048 00:39:39,260 --> 00:39:36,720 robust and this is seen across many many 1049 00:39:40,880 --> 00:39:39,270 models big decrease in the amount of 1050 00:39:44,480 --> 00:39:40,890 high clouds that we have with 1051 00:39:46,520 --> 00:39:44,490 aggregation so when we're clustered 1052 00:39:48,260 --> 00:39:46,530 we're warmer we're drier or especially 1053 00:39:50,930 --> 00:39:48,270 drier in the dry regions and we have 1054 00:39:52,070 --> 00:39:50,940 fewer clouds so we've changed the 1055 00:39:54,440 --> 00:39:52,080 temperature of the humidity in the 1056 00:39:56,210 --> 00:39:54,450 cloudiness and these are all things that 1057 00:39:59,390 --> 00:39:56,220 are really important for the radiative 1058 00:40:00,470 --> 00:39:59,400 energy budget okay so those changes have 1059 00:40:02,690 --> 00:40:00,480 big consequence 1060 00:40:04,550 --> 00:40:02,700 for the energy budget of our climate in 1061 00:40:06,920 --> 00:40:04,560 this case when the convection is more 1062 00:40:09,200 --> 00:40:06,930 clustered together so with aggregation 1063 00:40:10,940 --> 00:40:09,210 as the convection is more clustered we 1064 00:40:12,680 --> 00:40:10,950 have an increase in the outgoing 1065 00:40:15,920 --> 00:40:12,690 long-wave radiation this is again these 1066 00:40:17,690 --> 00:40:15,930 are all thoughts of time evolving in the 1067 00:40:19,790 --> 00:40:17,700 simulation so as time goes on things get 1068 00:40:21,650 --> 00:40:19,800 more clustered so we've got more 1069 00:40:24,950 --> 00:40:21,660 radiation escaping to space from the top 1070 00:40:26,800 --> 00:40:24,960 of the atmosphere and because we kind of 1071 00:40:29,150 --> 00:40:26,810 have these opposing changes and clouds 1072 00:40:31,220 --> 00:40:29,160 less high clouds but more low clouds 1073 00:40:33,050 --> 00:40:31,230 they reflected shorter radiation doesn't 1074 00:40:36,740 --> 00:40:33,060 really change much and so our next top 1075 00:40:39,349 --> 00:40:36,750 of atmosphere flux is also reduced in 1076 00:40:41,810 --> 00:40:39,359 magnitude so we have less flux into the 1077 00:40:44,060 --> 00:40:41,820 top faster so overall we're cooling more 1078 00:40:47,900 --> 00:40:44,070 okay there's more energy being lost to 1079 00:40:49,849 --> 00:40:47,910 space when we're aggregating that's 1080 00:40:52,310 --> 00:40:49,859 reflected as well by an increase in the 1081 00:40:54,410 --> 00:40:52,320 troposphere radiative cooling so this is 1082 00:40:55,940 --> 00:40:54,420 the column rate of cooling this is the 1083 00:40:56,990 --> 00:40:55,950 difference in fluxes from the top of 1084 00:40:59,450 --> 00:40:57,000 atmosphere or the surface or 1085 00:41:02,000 --> 00:40:59,460 equivalently the vertical integral of 1086 00:41:04,550 --> 00:41:02,010 the radiative pulling it's more negative 1087 00:41:06,589 --> 00:41:04,560 when we have convection clustered so 1088 00:41:08,059 --> 00:41:06,599 again we're cooling more when the 1089 00:41:10,099 --> 00:41:08,069 convection is tossed together and that's 1090 00:41:12,530 --> 00:41:10,109 because when it all comes together we've 1091 00:41:15,140 --> 00:41:12,540 opened up these huge areas that are dry 1092 00:41:17,540 --> 00:41:15,150 and clear and can radiate very strongly 1093 00:41:19,430 --> 00:41:17,550 to face the big you know radiator fins 1094 00:41:22,790 --> 00:41:19,440 as you know they put it in one of his 1095 00:41:24,260 --> 00:41:22,800 seminal papers of the atmosphere in 1096 00:41:25,880 --> 00:41:24,270 addition we have some changes to the 1097 00:41:27,500 --> 00:41:25,890 surface energy budget we have a decrease 1098 00:41:30,500 --> 00:41:27,510 in the energy gained by the surface and 1099 00:41:32,150 --> 00:41:30,510 because it's generally windier when 1100 00:41:35,089 --> 00:41:32,160 convection is clustered we have an 1101 00:41:37,819 --> 00:41:35,099 increase in surface enthalpy fluxes now 1102 00:41:39,319 --> 00:41:37,829 since a lot of this is evaporation we 1103 00:41:41,390 --> 00:41:39,329 have to have water balance in the system 1104 00:41:43,910 --> 00:41:41,400 so we have more evaporation we also have 1105 00:41:46,910 --> 00:41:43,920 more rainfall and so we see an increase 1106 00:41:48,410 --> 00:41:46,920 in mean precipitation so these are all 1107 00:41:50,839 --> 00:41:48,420 things that are really important for 1108 00:41:53,240 --> 00:41:50,849 sort of setting the climate of our 1109 00:41:55,370 --> 00:41:53,250 atmosphere and you know how strongly it 1110 00:41:58,069 --> 00:41:55,380 would it would respond to perturbations 1111 00:41:59,599 --> 00:41:58,079 so you can imagine you know if we have a 1112 00:42:01,339 --> 00:41:59,609 radiative forcing like we're saying 1113 00:42:03,200 --> 00:42:01,349 increasing carbon dioxide or something 1114 00:42:05,030 --> 00:42:03,210 and the amount of clustering of 1115 00:42:07,130 --> 00:42:05,040 convection changes then all of these 1116 00:42:09,800 --> 00:42:07,140 energy balances would change and that 1117 00:42:11,690 --> 00:42:09,810 would affect how sensitive our climate 1118 00:42:14,089 --> 00:42:11,700 would be you know potentially to a 1119 00:42:14,880 --> 00:42:14,099 climate perturbation and so this you 1120 00:42:16,740 --> 00:42:14,890 know 1121 00:42:19,099 --> 00:42:16,750 potentially big implications which as a 1122 00:42:21,720 --> 00:42:19,109 field we're still trying to figure out 1123 00:42:24,539 --> 00:42:21,730 so just to summarize self aggregation 1124 00:42:26,849 --> 00:42:24,549 warms and drives the mean state reduces 1125 00:42:28,559 --> 00:42:26,859 high clouds enhances the dryness of dry 1126 00:42:30,630 --> 00:42:28,569 regions increases the ability of the 1127 00:42:32,730 --> 00:42:30,640 atmosphere to cool to face I didn't talk 1128 00:42:34,700 --> 00:42:32,740 about this but it might be temperature 1129 00:42:36,990 --> 00:42:34,710 dependent again still figuring that out 1130 00:42:40,440 --> 00:42:37,000 and so therefore it might be important 1131 00:42:42,390 --> 00:42:40,450 for climate so in the last minute or two 1132 00:42:44,759 --> 00:42:42,400 it's want to kind of conclude here by 1133 00:42:46,470 --> 00:42:44,769 kind of reminding you that in reg 1134 00:42:48,450 --> 00:42:46,480 radiant convective equilibrium this kind 1135 00:42:50,730 --> 00:42:48,460 of idealized state of the tropical 1136 00:42:52,920 --> 00:42:50,740 atmosphere the organization of 1137 00:42:54,960 --> 00:42:52,930 convection takes the form of this self 1138 00:42:57,630 --> 00:42:54,970 aggregation which is the spontaneous 1139 00:43:00,720 --> 00:42:57,640 clustering of convection in an otherwise 1140 00:43:03,059 --> 00:43:00,730 homogeneous environment where we go from 1141 00:43:05,759 --> 00:43:03,069 scattered convection in the beginning 1142 00:43:07,529 --> 00:43:05,769 towards clustered at the end and that's 1143 00:43:09,210 --> 00:43:07,539 driven by these feedbacks involving 1144 00:43:12,450 --> 00:43:09,220 radiation and surface fluxes and how 1145 00:43:14,670 --> 00:43:12,460 they interact with convection so this 1146 00:43:16,380 --> 00:43:14,680 has all been you know simulations in 1147 00:43:17,549 --> 00:43:16,390 this idealized world a radiant 1148 00:43:19,259 --> 00:43:17,559 convective equilibrium where we have 1149 00:43:20,700 --> 00:43:19,269 this nice statistical balance between 1150 00:43:23,160 --> 00:43:20,710 the convective heating and the radiative 1151 00:43:24,839 --> 00:43:23,170 cooling and as I said that works on 1152 00:43:26,849 --> 00:43:24,849 average as a representation of our 1153 00:43:29,220 --> 00:43:26,859 atmosphere but I didn't do any 1154 00:43:31,589 --> 00:43:29,230 individual spot it doesn't hold and the 1155 00:43:32,819 --> 00:43:31,599 models configuration that in these 1156 00:43:34,650 --> 00:43:32,829 simulations I've been talking about is 1157 00:43:37,380 --> 00:43:34,660 pretty different from how the real earth 1158 00:43:39,359 --> 00:43:37,390 looks so a question that I get asked a 1159 00:43:42,029 --> 00:43:39,369 lot and that I think about a lot is how 1160 00:43:45,749 --> 00:43:42,039 my self aggregation manifest in the real 1161 00:43:48,029 --> 00:43:45,759 world right so in rce self aggregation 1162 00:43:49,799 --> 00:43:48,039 is the RET manifestation of these rated 1163 00:43:53,249 --> 00:43:49,809 connected feedbacks but our seee world 1164 00:43:57,539 --> 00:43:53,259 looks pretty different than earth okay 1165 00:43:59,789 --> 00:43:57,549 on earth there's rotation for one 1166 00:44:02,069 --> 00:43:59,799 there's gradients and temperature ray 1167 00:44:05,670 --> 00:44:02,079 it's warmer near the equator than it is 1168 00:44:07,710 --> 00:44:05,680 at the pole there's land and there's all 1169 00:44:09,029 --> 00:44:07,720 sorts of dynamical disturbances and 1170 00:44:10,859 --> 00:44:09,039 other circulations caused by other 1171 00:44:12,180 --> 00:44:10,869 things mid-latitude weather systems the 1172 00:44:13,410 --> 00:44:12,190 Hadley circulation Walker circulation 1173 00:44:16,140 --> 00:44:13,420 all that sort of stuff 1174 00:44:18,089 --> 00:44:16,150 so you know there's a lot of other 1175 00:44:19,680 --> 00:44:18,099 things that can control the decision of 1176 00:44:22,980 --> 00:44:19,690 convection in the real world and cause 1177 00:44:25,440 --> 00:44:22,990 it to organize and you know is this is 1178 00:44:26,609 --> 00:44:25,450 this meaningful basically is is self egg 1179 00:44:28,500 --> 00:44:26,619 region actually happening in the world 1180 00:44:30,570 --> 00:44:28,510 world well 1181 00:44:33,840 --> 00:44:30,580 we don't really truly know the answer to 1182 00:44:36,150 --> 00:44:33,850 that but what I think is that the same 1183 00:44:38,130 --> 00:44:36,160 feedbacks these radiative convective 1184 00:44:39,800 --> 00:44:38,140 surface flux feedbacks that cause self 1185 00:44:42,810 --> 00:44:39,810 aggregation in radiative equilibrium 1186 00:44:45,960 --> 00:44:42,820 those same feedbacks may contribute on 1187 00:44:48,210 --> 00:44:45,970 real earth to help convection that's 1188 00:44:51,090 --> 00:44:48,220 maybe initially organized by something 1189 00:44:54,480 --> 00:44:51,100 else help that stay organized or help it 1190 00:44:56,670 --> 00:44:54,490 organize more quickly I think that those 1191 00:44:58,500 --> 00:44:56,680 same feedbacks can help tropical 1192 00:45:00,870 --> 00:44:58,510 cyclones which are one special type of 1193 00:45:04,470 --> 00:45:00,880 convective organization help them form 1194 00:45:06,510 --> 00:45:04,480 and intensify and still in particular 1195 00:45:09,030 --> 00:45:06,520 there's growing evidence that cloud 1196 00:45:10,859 --> 00:45:09,040 radiative feedbacks specifically that 1197 00:45:12,720 --> 00:45:10,869 are driving self aggregation those cloud 1198 00:45:15,050 --> 00:45:12,730 rated feedbacks might also be very 1199 00:45:17,700 --> 00:45:15,060 important for hurricane and typhoon 1200 00:45:20,280 --> 00:45:17,710 formation and we already know they also 1201 00:45:23,190 --> 00:45:20,290 affect other things about its structure 1202 00:45:25,410 --> 00:45:23,200 and it's also possible that these 1203 00:45:27,090 --> 00:45:25,420 radiative convective feedbacks that give 1204 00:45:29,220 --> 00:45:27,100 us self aggregation they in the real 1205 00:45:31,500 --> 00:45:29,230 world when we're on a equatorial 1206 00:45:33,300 --> 00:45:31,510 rotating data plane might give rise to 1207 00:45:36,390 --> 00:45:33,310 the Matt and Julian oscillation this 1208 00:45:38,070 --> 00:45:36,400 planetary scale progression of a clump 1209 00:45:40,020 --> 00:45:38,080 of convection from the Indian to Pacific 1210 00:45:42,240 --> 00:45:40,030 Ocean and so there's growing evidence 1211 00:45:44,780 --> 00:45:42,250 that the mjo which is pictured here 1212 00:45:47,370 --> 00:45:44,790 again it's just this big blob of clouds 1213 00:45:51,090 --> 00:45:47,380 that the MgO is what's called a moisture 1214 00:45:53,220 --> 00:45:51,100 mode that is destabilized by these cloud 1215 00:45:56,520 --> 00:45:53,230 radiative feedbacks and indeed if you 1216 00:45:58,109 --> 00:45:56,530 take a model that was initially RCE that 1217 00:46:01,040 --> 00:45:58,119 self aggravated and you stick it on a 1218 00:46:03,750 --> 00:46:01,050 beta plane so we have rotation that 1219 00:46:05,810 --> 00:46:03,760 varies with latitude then we get 1220 00:46:08,970 --> 00:46:05,820 something that kind of looks like an MJ 1221 00:46:11,670 --> 00:46:08,980 so that those are the areas where I 1222 00:46:13,740 --> 00:46:11,680 think you know this self aggregation 1223 00:46:17,070 --> 00:46:13,750 phenomena is playing out in the real 1224 00:46:18,750 --> 00:46:17,080 atmosphere as well as just kind of 1225 00:46:22,050 --> 00:46:18,760 helping convection and clump together 1226 00:46:23,190 --> 00:46:22,060 more generally and a lot of my work now 1227 00:46:25,500 --> 00:46:23,200 is you know trying to make these 1228 00:46:27,690 --> 00:46:25,510 connections stronger so what I'd like to 1229 00:46:29,130 --> 00:46:27,700 leave you with then is something for you 1230 00:46:31,380 --> 00:46:29,140 guys to think about and maybe chat with 1231 00:46:33,359 --> 00:46:31,390 me about at the break is could this 1232 00:46:36,960 --> 00:46:33,369 convective cell of aggregation occur in 1233 00:46:38,670 --> 00:46:36,970 exoplanet atmospheres I mean to me it's 1234 00:46:41,220 --> 00:46:38,680 super exciting because with exoplanets I 1235 00:46:42,070 --> 00:46:41,230 mean there's such a huge diversity of 1236 00:46:44,050 --> 00:46:42,080 possible 1237 00:46:45,700 --> 00:46:44,060 planets and atmospheres out there maybe 1238 00:46:47,860 --> 00:46:45,710 there's one where it looks like RCE 1239 00:46:49,750 --> 00:46:47,870 where it's an aqua planet and it's so 1240 00:46:50,920 --> 00:46:49,760 far away from the Sun you don't have you 1241 00:46:52,510 --> 00:46:50,930 know Meridian 1242 00:46:55,210 --> 00:46:52,520 insulation gradients and it rotates 1243 00:46:57,880 --> 00:46:55,220 really slowly or something like that in 1244 00:46:59,770 --> 00:46:57,890 general in order for this to happen you 1245 00:47:01,420 --> 00:46:59,780 need to have an atmosphere where you 1246 00:47:03,040 --> 00:47:01,430 have convection happening in the 1247 00:47:04,870 --> 00:47:03,050 atmosphere so it needs to be unstable in 1248 00:47:07,690 --> 00:47:04,880 a way to have convection and that 1249 00:47:10,630 --> 00:47:07,700 convection would need to occur in a way 1250 00:47:13,180 --> 00:47:10,640 that the it affects the distribution of 1251 00:47:16,450 --> 00:47:13,190 a radiative lee active condensable 1252 00:47:18,790 --> 00:47:16,460 species and it needs to do that in a way 1253 00:47:20,380 --> 00:47:18,800 such that you have less radiative 1254 00:47:22,330 --> 00:47:20,390 cooling where you would have the 1255 00:47:24,370 --> 00:47:22,340 convection which then invigorates it so 1256 00:47:25,480 --> 00:47:24,380 you guys have to tell me about what 1257 00:47:27,220 --> 00:47:25,490 convection looks like and what the 1258 00:47:29,650 --> 00:47:27,230 condensable species are and if it's 1259 00:47:31,420 --> 00:47:29,660 radiation behaves in that way but I 1260 00:47:33,400 --> 00:47:31,430 think that there's some exciting 1261 00:47:34,420 --> 00:47:33,410 possibilities there and hopefully this 1262 00:47:37,270 --> 00:47:34,430 has been interesting for you to hear 1263 00:47:38,500 --> 00:47:37,280 about and think about maybe change the 1264 00:47:40,810 --> 00:47:38,510 way you think about how convection 1265 00:47:41,710 --> 00:47:40,820 behaves in atmosphere so thank you very 1266 00:48:03,290 --> 00:47:41,720 much for your attention 1267 00:48:11,940 --> 00:48:07,290 so oh I'm Doreen Abbott from University 1268 00:48:14,940 --> 00:48:11,950 of Chicago in the in the small scale in 1269 00:48:18,180 --> 00:48:14,950 the doubly periodic domains it looked 1270 00:48:21,570 --> 00:48:18,190 like mostly the convection aggregates at 1271 00:48:23,010 --> 00:48:21,580 the domain scale in the radiative 1272 00:48:25,440 --> 00:48:23,020 convective equilibrium global 1273 00:48:27,090 --> 00:48:25,450 simulations there seemed to be a scale 1274 00:48:30,210 --> 00:48:27,100 and it looked like in the different 1275 00:48:33,150 --> 00:48:30,220 models the scale was different what sets 1276 00:48:36,330 --> 00:48:33,160 that scale that's a good question 1277 00:48:37,560 --> 00:48:36,340 we don't know the answer certainly in 1278 00:48:39,180 --> 00:48:37,570 these as you mentioned any sort of 1279 00:48:41,940 --> 00:48:39,190 square things it seems to scale with a 1280 00:48:43,440 --> 00:48:41,950 domain size so you have to people have 1281 00:48:45,210 --> 00:48:43,450 finally figured out if you get up to ten 1282 00:48:46,650 --> 00:48:45,220 thousand kilometers then finally you 1283 00:48:48,300 --> 00:48:46,660 start getting a multiple clusters so 1284 00:48:50,340 --> 00:48:48,310 there does seem to be some natural 1285 00:48:52,680 --> 00:48:50,350 limiting length scale to the size of 1286 00:48:54,330 --> 00:48:52,690 these clusters this is on the order of a 1287 00:48:56,520 --> 00:48:54,340 couple thousand kilometers like order a 1288 00:48:58,590 --> 00:48:56,530 thousand kilometers you would say but as 1289 00:49:02,250 --> 00:48:58,600 for what physically sets that we don't 1290 00:49:04,650 --> 00:49:02,260 know I had a theory sort of involving 1291 00:49:06,690 --> 00:49:04,660 boundary layer and moistening lengths 1292 00:49:08,640 --> 00:49:06,700 but it kind of worked it kind of didn't 1293 00:49:10,500 --> 00:49:08,650 so I don't really believe it people are 1294 00:49:12,840 --> 00:49:10,510 still working on it 1295 00:49:14,940 --> 00:49:12,850 but it's sort of it's tricky it's a 1296 00:49:18,150 --> 00:49:14,950 tricky problem because in a non rotating 1297 00:49:19,710 --> 00:49:18,160 state we lose a lot of our natural 1298 00:49:21,660 --> 00:49:19,720 length scales in the atmosphere there's 1299 00:49:22,650 --> 00:49:21,670 no rafi deformation radius or things 1300 00:49:24,270 --> 00:49:22,660 like that there's maybe some length 1301 00:49:25,470 --> 00:49:24,280 scales involving you know frictional 1302 00:49:27,360 --> 00:49:25,480 dissipation or something like that 1303 00:49:30,390 --> 00:49:27,370 so we're still trying to figure it out 1304 00:49:31,890 --> 00:49:30,400 but you know as for I think it's you 1305 00:49:34,080 --> 00:49:31,900 know if it's set by some you know 1306 00:49:35,670 --> 00:49:34,090 fundamental property all these models 1307 00:49:37,410 --> 00:49:35,680 are you know having different treatments 1308 00:49:39,300 --> 00:49:37,420 of boundary layer representations 1309 00:49:40,860 --> 00:49:39,310 convection etc so I think it's 1310 00:49:43,800 --> 00:49:40,870 reasonable that they would do that give 1311 00:49:51,720 --> 00:49:43,810 it a different length but that's an open 1312 00:49:53,970 --> 00:49:51,730 area research hi Ralph Lorenz Applied 1313 00:49:56,670 --> 00:49:53,980 Physics lab a really interesting 1314 00:49:59,520 --> 00:49:56,680 perspective on eager to sort see this 1315 00:50:01,120 --> 00:49:59,530 apply to solar system 1316 00:50:04,359 --> 00:50:01,130 Emma spheres as well as actually 1317 00:50:06,249 --> 00:50:04,369 planets and I'm grateful to to Torian 1318 00:50:08,079 --> 00:50:06,259 for asking about the length scale and 1319 00:50:10,150 --> 00:50:08,089 periodicity excited the same same 1320 00:50:12,150 --> 00:50:10,160 question the other question I had was 1321 00:50:16,180 --> 00:50:12,160 with respect to the the long wave 1322 00:50:20,499 --> 00:50:16,190 feedback where the slide said subsidence 1323 00:50:22,240 --> 00:50:20,509 causes low low-level cloud and I wasn't 1324 00:50:24,549 --> 00:50:22,250 sure how that worked is that just 1325 00:50:26,170 --> 00:50:24,559 substance causes cloud to occur at a 1326 00:50:29,079 --> 00:50:26,180 lower level than it would otherwise do 1327 00:50:30,839 --> 00:50:29,089 or I mean how does how does subsidence 1328 00:50:33,339 --> 00:50:30,849 cause condensation that's 1329 00:50:36,160 --> 00:50:33,349 counterintuitive well so it's more that 1330 00:50:37,930 --> 00:50:36,170 it's affecting is sort of the profile of 1331 00:50:41,049 --> 00:50:37,940 cooling and heating in the atmosphere so 1332 00:50:42,490 --> 00:50:41,059 you would have any convection that would 1333 00:50:44,470 --> 00:50:42,500 happen would be captain couldn't get any 1334 00:50:46,599 --> 00:50:44,480 higher and in the real tropical 1335 00:50:48,460 --> 00:50:46,609 atmosphere we generally see in 1336 00:50:50,529 --> 00:50:48,470 subsidence regions in the low levels you 1337 00:50:52,660 --> 00:50:50,539 have a trade inversion and so below that 1338 00:50:54,160 --> 00:50:52,670 the air is unstable and you can have 1339 00:50:56,109 --> 00:50:54,170 right at the top of that layer cloud 1340 00:50:59,079 --> 00:50:56,119 formation but it can't go any higher and 1341 00:51:02,140 --> 00:50:59,089 so having subsidence actually helps keep 1342 00:51:05,289 --> 00:51:02,150 it kind of confined to that layer and 1343 00:51:06,730 --> 00:51:05,299 that and then again the strong cooling 1344 00:51:08,650 --> 00:51:06,740 that you have at the top of that layer 1345 00:51:10,120 --> 00:51:08,660 that helps maintain the convection that 1346 00:51:11,440 --> 00:51:10,130 you're kind of you can get convection if 1347 00:51:12,940 --> 00:51:11,450 you're heating from below or cooling 1348 00:51:15,490 --> 00:51:12,950 from above and so both of those are 1349 00:51:21,789 --> 00:51:15,500 contributing to the local low clouds 1350 00:51:24,279 --> 00:51:21,799 there yeah Eric I don't see numbers 1351 00:51:26,259 --> 00:51:24,289 Hawaii so my my question is a segue to 1352 00:51:28,089 --> 00:51:26,269 dorians which is basically you know 1353 00:51:30,460 --> 00:51:28,099 these simulations are idealized in terms 1354 00:51:33,430 --> 00:51:30,470 of their surface by necessity but have 1355 00:51:36,849 --> 00:51:33,440 people looked at or you looked at the 1356 00:51:40,509 --> 00:51:36,859 fragility of the of the process to an 1357 00:51:42,940 --> 00:51:40,519 imposed physical structure no the real 1358 00:51:44,859 --> 00:51:42,950 earth has lakes mountain ranges force 1359 00:51:47,499 --> 00:51:44,869 lots of lots of air that's a 1360 00:51:49,990 --> 00:51:47,509 heterogeneity right and do those play 1361 00:51:52,690 --> 00:51:50,000 into basically setting for example the 1362 00:51:54,460 --> 00:51:52,700 length scale of this aggregation this is 1363 00:51:57,309 --> 00:51:54,470 something that might distinguish for 1364 00:51:59,799 --> 00:51:57,319 example an aqua planet from from a 1365 00:52:02,440 --> 00:51:59,809 planet which has no such smooth surface 1366 00:52:04,839 --> 00:52:02,450 yeah so so if you put in surface 1367 00:52:09,190 --> 00:52:04,849 heterogeneities that kind of tends to 1368 00:52:11,079 --> 00:52:09,200 tell the convection where to cluster so 1369 00:52:13,299 --> 00:52:11,089 for example if we put in something kind 1370 00:52:14,920 --> 00:52:13,309 of like an island in these simulations 1371 00:52:16,359 --> 00:52:14,930 you'll have a 1372 00:52:18,190 --> 00:52:16,369 cluster over the island but then you 1373 00:52:20,290 --> 00:52:18,200 might have other convective aggregated 1374 00:52:21,940 --> 00:52:20,300 clusters elsewhere as well 1375 00:52:23,829 --> 00:52:21,950 people have done simulations also with 1376 00:52:25,240 --> 00:52:23,839 interactive surface temperatures which 1377 00:52:27,430 --> 00:52:25,250 can cause some interesting kind of 1378 00:52:30,670 --> 00:52:27,440 internal variability in the structure of 1379 00:52:33,220 --> 00:52:30,680 the aggregation but yeah I mean I guess 1380 00:52:35,650 --> 00:52:33,230 basically any any heterogeneity is gonna 1381 00:52:37,359 --> 00:52:35,660 probably be a stronger influence on 1382 00:52:41,200 --> 00:52:37,369 telling the convection specifically 1383 00:52:43,900 --> 00:52:41,210 where to organize then then this random 1384 00:52:46,990 --> 00:52:43,910 kind of self aggregation process and so 1385 00:52:48,549 --> 00:52:47,000 that's why you know it's sort of you 1386 00:52:50,020 --> 00:52:48,559 have to kind of think a lot about how 1387 00:52:51,700 --> 00:52:50,030 this is happening in the real world I 1388 00:52:53,589 --> 00:52:51,710 mean there are places in the tropical 1389 00:52:56,589 --> 00:52:53,599 ocean where there are relatively weak 1390 00:52:58,510 --> 00:52:56,599 sea surface temperature gradients but in 1391 00:52:59,920 --> 00:52:58,520 you know kind of land areas and things 1392 00:53:01,780 --> 00:52:59,930 like that you know those are probably 1393 00:53:03,970 --> 00:53:01,790 stronger influences 1394 00:53:06,010 --> 00:53:03,980 if for example you put in a sea surface 1395 00:53:08,680 --> 00:53:06,020 temperature like a Mach Walker kind of 1396 00:53:10,329 --> 00:53:08,690 circulation type thing again the 1397 00:53:12,010 --> 00:53:10,339 convection kind of chooses to be over 1398 00:53:13,750 --> 00:53:12,020 the warm areas but these sort of 1399 00:53:21,280 --> 00:53:13,760 feedbacks can still be going on to 1400 00:53:24,450 --> 00:53:21,290 amplify the organization that evolves my 1401 00:53:28,059 --> 00:53:24,460 question just actually got it 1402 00:53:30,130 --> 00:53:28,069 why does self aggregation saturate like 1403 00:53:32,789 --> 00:53:30,140 what negative feedback kicks in at the 1404 00:53:38,170 --> 00:53:32,799 end that makes it fly it makes it stop 1405 00:53:39,910 --> 00:53:38,180 well I mean you have to you have to have 1406 00:53:41,799 --> 00:53:39,920 some amount of convection happening 1407 00:53:42,819 --> 00:53:41,809 right because we have cooling happening 1408 00:53:45,190 --> 00:53:42,829 so we have to have convection of 1409 00:53:47,020 --> 00:53:45,200 Ballantyne so you can't have the cluster 1410 00:53:50,410 --> 00:53:47,030 you know collapse into a pink point and 1411 00:53:52,780 --> 00:53:50,420 go to zero and so in terms of what scale 1412 00:53:54,250 --> 00:53:52,790 takes it to do that that goes back to 1413 00:53:56,349 --> 00:53:54,260 the question the Dorian asked about the 1414 00:53:57,069 --> 00:53:56,359 length scale you know why does do we 1415 00:53:59,109 --> 00:53:57,079 stop 1416 00:54:00,339 --> 00:53:59,119 you know seeing an increase in the the 1417 00:54:02,559 --> 00:54:00,349 variance and the drying of the dry 1418 00:54:05,170 --> 00:54:02,569 regions I mean it is adjusting to a new 1419 00:54:06,730 --> 00:54:05,180 equilibrium basically and once it gets 1420 00:54:08,980 --> 00:54:06,740 into a state where it can rebalance 1421 00:54:11,260 --> 00:54:08,990 things out and have this energy balance 1422 00:54:14,500 --> 00:54:11,270 then it's happy and then it stays there 1423 00:54:17,260 --> 00:54:14,510 and I one thing I didn't mention is that 1424 00:54:19,000 --> 00:54:17,270 once it's aggregates it stays that way 1425 00:54:20,799 --> 00:54:19,010 as long as they run the simulation 1426 00:54:22,480 --> 00:54:20,809 basically there's a lot of hysteresis in 1427 00:54:24,579 --> 00:54:22,490 the process and in fact if you 1428 00:54:26,380 --> 00:54:24,589 initialize a simulation from an 1429 00:54:27,820 --> 00:54:26,390 aggregated state even in conditions 1430 00:54:29,980 --> 00:54:27,830 where it might not organize 1431 00:54:31,870 --> 00:54:29,990 on its own it's it's quite happy to stay 1432 00:54:34,510 --> 00:54:31,880 organized and so you have to do a lot to 1433 00:54:37,000 --> 00:54:34,520 break it up the easiest way to break it 1434 00:54:39,400 --> 00:54:37,010 up is to put in vertical wind shear into 1435 00:54:42,550 --> 00:54:39,410 the simulation because that's gonna 1436 00:54:44,140 --> 00:54:42,560 start causing mixing between these moist 1437 00:54:46,210 --> 00:54:44,150 and dry regions so anything that gives 1438 00:54:49,330 --> 00:54:46,220 you that might break it up but otherwise 1439 00:54:51,820 --> 00:54:49,340 it's just so and the areas outside the 1440 00:54:53,800 --> 00:54:51,830 convective cluster it's so unfavorable 1441 00:54:55,390 --> 00:54:53,810 to convection there that it just stays 1442 00:55:00,790 --> 00:54:55,400 it's impose its own little happy 1443 00:55:02,530 --> 00:55:00,800 environment John Harington University of 1444 00:55:06,820 --> 00:55:02,540 Central Florida welcome to Florida where 1445 00:55:09,270 --> 00:55:06,830 are you so this is just another step 1446 00:55:13,360 --> 00:55:09,280 back question the more I hear about 1447 00:55:15,730 --> 00:55:13,370 short-term processes like convection and 1448 00:55:17,680 --> 00:55:15,740 waves and so forth the more I'm 1449 00:55:21,190 --> 00:55:17,690 convinced that they're critically 1450 00:55:23,080 --> 00:55:21,200 important to long-term processes and I 1451 00:55:25,830 --> 00:55:23,090 just wanted to ask I know this is a very 1452 00:55:29,290 --> 00:55:25,840 naive question but how do you connect 1453 00:55:30,880 --> 00:55:29,300 processes like this to say a climate 1454 00:55:33,970 --> 00:55:30,890 model that's gonna run for a billion 1455 00:55:36,430 --> 00:55:33,980 years that's a good question I mean I I 1456 00:55:38,410 --> 00:55:36,440 think the connection is through how 1457 00:55:40,390 --> 00:55:38,420 these smaller scale processes kind of 1458 00:55:42,040 --> 00:55:40,400 rectify on the larger scale like I was 1459 00:55:44,230 --> 00:55:42,050 talking about towards the end with these 1460 00:55:47,980 --> 00:55:44,240 impacts of a larger scale climate I mean 1461 00:55:50,650 --> 00:55:47,990 if you know if if the atmosphere you 1462 00:55:52,330 --> 00:55:50,660 know was totally the same regardless of 1463 00:55:55,030 --> 00:55:52,340 whether convection was in a cluster or a 1464 00:55:57,310 --> 00:55:55,040 line or a circle or a star or whatever 1465 00:55:59,320 --> 00:55:57,320 then maybe maybe those smell scale 1466 00:56:02,740 --> 00:55:59,330 processes wouldn't matter for a billion 1467 00:56:05,140 --> 00:56:02,750 year simulation but you know if any 1468 00:56:07,360 --> 00:56:05,150 process sort of has an impact that 1469 00:56:10,210 --> 00:56:07,370 persists over a longer time period than 1470 00:56:11,920 --> 00:56:10,220 its lifetime or on a larger scale of it 1471 00:56:14,020 --> 00:56:11,930 then it's something that would be 1472 00:56:15,850 --> 00:56:14,030 relevant over there's long scales from a 1473 00:56:18,760 --> 00:56:15,860 kind of more fundamental perspective you 1474 00:56:21,010 --> 00:56:18,770 know convection itself you know it's 1475 00:56:23,230 --> 00:56:21,020 setting the structure of the temperature 1476 00:56:25,090 --> 00:56:23,240 profiles in the atmosphere I mean the 1477 00:56:26,860 --> 00:56:25,100 reason you know I have the surface 1478 00:56:27,910 --> 00:56:26,870 temperature that we do and the rate of 1479 00:56:29,620 --> 00:56:27,920 decrease of temperature with height 1480 00:56:31,570 --> 00:56:29,630 that's because of these small-scale 1481 00:56:33,100 --> 00:56:31,580 hour-long you know time period 1482 00:56:35,440 --> 00:56:33,110 convection things and so that's you know 1483 00:56:37,540 --> 00:56:35,450 the strongest example for how these 1484 00:56:40,330 --> 00:56:37,550 small-scale processes are something that 1485 00:56:41,130 --> 00:56:40,340 we have to take into account so you know 1486 00:56:43,170 --> 00:56:41,140 it's 1487 00:56:44,910 --> 00:56:43,180 that's why climate models where they're 1488 00:56:46,640 --> 00:56:44,920 parameterizing these things have to work 1489 00:56:49,170 --> 00:56:46,650 so hard to get those things right 1490 00:56:51,299 --> 00:56:49,180 because those small-scale processes do 1491 00:56:53,940 --> 00:56:51,309 play a big role when we're talking about 1492 00:56:57,660 --> 00:56:53,950 you know kind of centennial scale future 1493 00:56:59,460 --> 00:56:57,670 climate projections the biggest 1494 00:57:01,620 --> 00:56:59,470 uncertainty in those is how these 1495 00:57:03,240 --> 00:57:01,630 convective clouds are handled because of 1496 00:57:05,460 --> 00:57:03,250 some of these regions 1497 00:57:07,500 --> 00:57:05,470 you no longer scales billions of years I 1498 00:57:08,910 --> 00:57:07,510 mean again that's where you start having 1499 00:57:10,620 --> 00:57:08,920 other things like you know geologic 1500 00:57:13,079 --> 00:57:10,630 processes and carbon cycles starting to 1501 00:57:13,980 --> 00:57:13,089 become really important but you know we 1502 00:57:15,900 --> 00:57:13,990 still need to know what the temperature