WEBVTT 00:00:00.000 --> 00:00:01.997 [SQUEAKING] 00:00:01.997 --> 00:00:04.492 [RUSTLING] 00:00:04.492 --> 00:00:06.488 [CLICKING] 00:00:10.960 --> 00:00:13.480 R. SCOTT KEMP: We're going to figure out how many people 00:00:13.480 --> 00:00:15.060 nuclear power kills. 00:00:17.960 --> 00:00:20.440 I mean, we need an answer to this. 00:00:20.440 --> 00:00:26.380 So this is part of the externalities calculation. 00:00:26.380 --> 00:00:29.240 If you recall, we spent the first major part 00:00:29.240 --> 00:00:34.600 of the class working out the explicit internalized economics. 00:00:34.600 --> 00:00:38.000 But I said that that's not the whole story. 00:00:38.000 --> 00:00:40.720 We need also to have externalities. 00:00:40.720 --> 00:00:42.163 One externality is climate change. 00:00:42.163 --> 00:00:44.580 We talked about the difficulty of coming up with a number. 00:00:44.580 --> 00:00:46.520 So we said, well, we'll take the view 00:00:46.520 --> 00:00:49.640 that we don't know what it is, but we 00:00:49.640 --> 00:00:52.560 will be able to estimate how big it would 00:00:52.560 --> 00:00:55.880 have to be to justify things. 00:00:55.880 --> 00:00:58.620 But the difficulty of that is that 00:00:58.620 --> 00:01:03.460 under equal decarbonization, wind and solar mostly 00:01:03.460 --> 00:01:08.860 outcompetes nuclear, except for in the thermal markets where 00:01:08.860 --> 00:01:11.380 there's potentially some opportunity. 00:01:11.380 --> 00:01:14.180 And now the question is, is there an additional safety 00:01:14.180 --> 00:01:20.060 externality associated with nuclear, or other externalities 00:01:20.060 --> 00:01:24.620 associated with wind or solar or natural gas that actually change 00:01:24.620 --> 00:01:27.460 the ordering, if you will? 00:01:27.460 --> 00:01:30.164 We've done everything from explicit perspective. 00:01:30.160 --> 00:01:32.740 What about these externalities? 00:01:32.740 --> 00:01:35.740 What if it turns out that 00:01:35.740 --> 00:01:40.580 "Windmills are killing the birds." turns out to be a big problem? 00:01:40.580 --> 00:01:43.120 So we need to get into that. 00:01:43.120 --> 00:01:45.100 But the first thing we're going to do 00:01:45.100 --> 00:01:48.300 is calculate the externality for nuclear. 00:01:48.300 --> 00:01:53.900 So who has seen this plot? 00:01:53.900 --> 00:01:57.230 How many people feel they were-- 00:01:57.230 --> 00:01:58.710 yeah. 00:01:58.710 --> 00:02:01.030 How many people feel that this plot actually 00:02:01.030 --> 00:02:03.890 influenced their choice to become nuclear engineers? 00:02:07.810 --> 00:02:08.310 Yes? 00:02:08.310 --> 00:02:09.110 No? 00:02:09.110 --> 00:02:12.140 There's a kind of a shaky head. 00:02:12.140 --> 00:02:16.110 One half person says this plot had influence on them. 00:02:16.110 --> 00:02:18.950 It's a nice plot. 00:02:18.950 --> 00:02:21.510 I hope it's right. 00:02:21.510 --> 00:02:22.450 We'll find out. 00:02:22.450 --> 00:02:23.490 That's what we're going to do today. 00:02:23.490 --> 00:02:25.350 We're going to figure out what this plot is. 00:02:25.350 --> 00:02:27.630 So let's go. 00:02:27.630 --> 00:02:33.150 So I showed you this slide Wednesday. 00:02:33.150 --> 00:02:36.390 This is some of the isotopes that 00:02:36.390 --> 00:02:41.150 come out of reactor accidents which have large activity. 00:02:41.150 --> 00:02:43.750 And we mentioned that cesium-- sorry, the European spelling-- 00:02:43.750 --> 00:02:48.870 cesium-137 and also cesium-134 is 00:02:48.870 --> 00:02:52.470 significant on the early days. 00:02:52.470 --> 00:02:57.090 And these other ones, they dominate hundreds of years. 00:02:57.090 --> 00:02:59.570 But this is a log plot here. 00:02:59.570 --> 00:03:03.890 And these are the two isotopes of principal interest-- 00:03:03.890 --> 00:03:06.850 cesium-134 and cesium-137. 00:03:06.850 --> 00:03:09.430 Iodine-131, which is not shown on this plot, 00:03:09.430 --> 00:03:11.090 is even more important. 00:03:11.090 --> 00:03:14.010 It has a very short half life. 00:03:14.010 --> 00:03:18.390 So it decays entirely in a few months, a couple of years, 00:03:18.390 --> 00:03:19.410 basically. 00:03:19.410 --> 00:03:25.230 And it gives a big dose that we also have to take into account. 00:03:25.230 --> 00:03:29.590 But it is a different dynamic because it is short-lived. 00:03:29.590 --> 00:03:31.850 It's what you eat and breathe immediately 00:03:31.850 --> 00:03:33.290 following the accident. 00:03:33.290 --> 00:03:35.250 This other stuff, cesium just hangs around 00:03:35.250 --> 00:03:40.610 for centuries and just increases slowly the background radiation 00:03:40.610 --> 00:03:41.530 dose. 00:03:41.530 --> 00:03:45.362 So we are going to do the calculation for cesium. 00:03:45.362 --> 00:03:47.070 We'll look at some of the other isotopes. 00:03:47.070 --> 00:03:49.450 We'll add in iodine, but we won't do the calculation 00:03:49.450 --> 00:03:50.930 explicitly. 00:03:50.930 --> 00:03:54.660 And we'll just try to take into account the biggest and most 00:03:54.660 --> 00:03:58.860 important isotopes and get a measure of the dose. 00:03:58.860 --> 00:04:02.980 Just as a mention, you've seen the periodic table before. 00:04:02.980 --> 00:04:07.940 Why is cesium-- and another one people hear about strontium-90? 00:04:07.940 --> 00:04:10.620 Why is strontium and caesium important? 00:04:10.620 --> 00:04:13.020 Well, because of the valence electrons 00:04:13.020 --> 00:04:15.800 are similar to in a case of strontium like calcium. 00:04:15.800 --> 00:04:18.380 So it sits in our bones. 00:04:18.380 --> 00:04:20.720 It's not a big beta-- it's not a big external emitter. 00:04:20.720 --> 00:04:22.880 If you're around strontium contamination, 00:04:22.880 --> 00:04:23.800 it's not a problem. 00:04:23.800 --> 00:04:25.220 But if you eat it, it's a problem 00:04:25.220 --> 00:04:28.700 because it then gets taken into your bones. 00:04:28.700 --> 00:04:32.820 And so like the history of above ground 00:04:32.820 --> 00:04:35.508 nuclear weapons testing was that we 00:04:35.508 --> 00:04:37.300 were testing all these nuclear bombs, which 00:04:37.300 --> 00:04:41.100 are little mini fission reactors, which 00:04:41.100 --> 00:04:44.540 disperse all of their fission products into the environment. 00:04:44.540 --> 00:04:48.260 And we were just-- the US government 00:04:48.260 --> 00:04:49.900 and other governments were doing this. 00:04:49.900 --> 00:04:55.600 And then it turned out that the strontium was emplacing itself 00:04:55.600 --> 00:04:58.040 in the teeth of babies. 00:04:58.040 --> 00:04:59.960 Instead of calcium apatite, which 00:04:59.960 --> 00:05:03.340 is the enamel coating of your teeth, it was strontium apatite. 00:05:03.340 --> 00:05:05.560 And you could measure the radioactivity 00:05:05.560 --> 00:05:08.560 of the teeth of newborn baby-- of young babies. 00:05:08.560 --> 00:05:14.080 And this fact caused essentially a political movement 00:05:14.080 --> 00:05:16.640 that forced nuclear weapons testing underground. 00:05:16.640 --> 00:05:21.413 So, in fact, the amount of fission products released 00:05:21.413 --> 00:05:23.080 from typical nuclear weapons is actually 00:05:23.080 --> 00:05:27.660 a lot smaller than what we would get from a large accident. 00:05:27.660 --> 00:05:31.340 So strontium does matter, mainly for ingestion pathways. 00:05:31.340 --> 00:05:33.800 We're not going to talk about it. 00:05:33.800 --> 00:05:35.960 Cesium looks like potassium. 00:05:35.960 --> 00:05:39.280 And so potassium is used everywhere in your body. 00:05:39.280 --> 00:05:45.220 So it just filters through all of your organs and stays around. 00:05:45.220 --> 00:05:47.920 So let's look at some of the source terms. 00:05:47.920 --> 00:05:50.860 We're going to look at two major accidents-- 00:05:50.860 --> 00:05:55.140 the two INES 7 events. 00:05:55.140 --> 00:05:58.220 This is an international nuclear event scale. 00:05:58.220 --> 00:05:59.620 It goes up to 7. 00:05:59.620 --> 00:06:00.453 7 is the worst. 00:06:00.453 --> 00:06:02.620 There has only been two of these events in history-- 00:06:02.620 --> 00:06:04.380 Chernobyl and Fukushima. 00:06:04.380 --> 00:06:07.420 So let's look at those events because they 00:06:07.420 --> 00:06:09.580 dominate the source. 00:06:09.580 --> 00:06:14.980 So Chernobyl was an RBMK-type water graphite reactor. 00:06:14.980 --> 00:06:17.980 It had no containment. 00:06:17.980 --> 00:06:21.200 They believed that it wasn't really possible. 00:06:21.200 --> 00:06:28.180 A steam explosion ejected fuel and melted the fuel 00:06:28.180 --> 00:06:29.760 and triggered fires. 00:06:29.760 --> 00:06:36.380 And the fires then helped to volatilize a lot of the fission 00:06:36.380 --> 00:06:40.340 products that were in the fuel that boil off at temperatures 00:06:40.340 --> 00:06:42.380 below the fire temperature. 00:06:42.380 --> 00:06:45.700 So there are still 11 of these reactors operating 00:06:45.700 --> 00:06:49.430 in the world today, all in the former Soviet Union. 00:06:49.430 --> 00:06:54.420 So they continue to pose whatever threat they pose. 00:06:54.420 --> 00:06:56.670 AUDIENCE: [INAUDIBLE] as the Soviets did actually make 00:06:56.670 --> 00:06:58.270 some changes after Chernobyl. 00:06:58.270 --> 00:06:59.790 They're not exactly the same, 00:06:59.790 --> 00:07:03.470 R. SCOTT KEMP: Yes, but they still don't have containment. 00:07:03.470 --> 00:07:08.270 So this thing released 85 petabecquerels of cesium 137, 00:07:08.270 --> 00:07:14.750 4,000 petabecquerels of non-noble gas activity total. 00:07:14.750 --> 00:07:18.390 But remember, petabecquerel is just decays. 00:07:18.390 --> 00:07:21.530 It doesn't tell you how bad the decays are. 00:07:21.530 --> 00:07:24.110 It doesn't tell you how much energy is in those decays. 00:07:24.110 --> 00:07:28.230 So it's not-- you can't just scale this number. 00:07:28.230 --> 00:07:32.910 The number we want to use here is 85. 00:07:32.910 --> 00:07:37.790 This is the contamination as calculated by the French. 00:07:37.790 --> 00:07:44.270 I think this is IRSN across Europe. 00:07:44.270 --> 00:07:47.250 75% of the total release is estimated 00:07:47.250 --> 00:07:53.050 to have been deposited on land somewhere in the world. 00:07:53.050 --> 00:07:56.530 And this is a log scale, so it's a little bit tricky 00:07:56.530 --> 00:07:58.810 to try to interpret. 00:07:58.810 --> 00:08:00.690 Fortunately, if we wanted to measure 00:08:00.690 --> 00:08:05.290 the average contamination, say, in all of Europe, 00:08:05.290 --> 00:08:10.270 we could try to voxelize or pixelize this and figure it out. 00:08:10.270 --> 00:08:16.050 But this European Commission issued this report in 1998, 00:08:16.050 --> 00:08:17.610 and they did it for us. 00:08:17.610 --> 00:08:21.210 And their number is 7 kilobecquerels 00:08:21.210 --> 00:08:23.490 is the European spatial average contamination-- 00:08:23.490 --> 00:08:27.130 7 kilobecquerels of cesium-137 per meter squared. 00:08:27.130 --> 00:08:29.130 So that's good to know. 00:08:29.130 --> 00:08:32.890 Now let's make a little observation here. 00:08:32.890 --> 00:08:36.330 The first observation is that in the last class 00:08:36.330 --> 00:08:39.010 I defended the linear no threshold model. 00:08:39.010 --> 00:08:41.890 If the linear no threshold model is correct, 00:08:41.890 --> 00:08:49.670 then that means 10 millisieverts of dose to one person 00:08:49.670 --> 00:08:53.550 is the same as 1 millisieverts of dose to 10 people. 00:08:53.550 --> 00:08:58.710 And so that makes calculating the dose response really easy. 00:08:58.710 --> 00:09:03.510 We just basically need to estimate the average population 00:09:03.510 --> 00:09:04.890 times the average dose. 00:09:04.890 --> 00:09:07.390 And because it's all linear, it's roughly correct. 00:09:07.390 --> 00:09:09.990 And that's what we're going to do. 00:09:09.990 --> 00:09:13.270 Now, what we really care about is not the average land 00:09:13.270 --> 00:09:15.350 contamination. 00:09:15.350 --> 00:09:18.990 It's the average exposure to a person. 00:09:18.990 --> 00:09:22.750 And people aren't evenly spread across all of the land. 00:09:22.750 --> 00:09:29.790 People-- like Western Europe has a higher population density 00:09:29.790 --> 00:09:31.270 than Siberia. 00:09:31.270 --> 00:09:35.990 And so we need to maybe accommodate for that. 00:09:35.990 --> 00:09:39.550 A thing that has been raised is that Ukraine, which I've now 00:09:39.550 --> 00:09:43.600 outlined here for you in black, is sometimes called 00:09:43.600 --> 00:09:45.840 the bread bowl of Europe. 00:09:45.840 --> 00:09:49.100 A huge amount of grain is grown there, 00:09:49.100 --> 00:09:53.220 which uptakes cesium from the ground-- 00:09:53.220 --> 00:09:54.940 looks like potassium-- into the grain. 00:09:54.940 --> 00:09:58.120 And then you eat it, and you have an ingestion dose. 00:09:58.120 --> 00:10:01.460 So arguably, the European spatial average, 00:10:01.460 --> 00:10:07.360 which includes all of uneventful space, is probably too low. 00:10:07.360 --> 00:10:11.760 Since we're really in the domain of single digits of significance 00:10:11.760 --> 00:10:13.560 here, I'm just going to round this 00:10:13.560 --> 00:10:18.640 to 10, to take into account the fact that they grow 00:10:18.640 --> 00:10:21.580 a lot of grain here, and Europe is here, and so on and so forth. 00:10:21.580 --> 00:10:21.800 Yep. 00:10:21.800 --> 00:10:23.342 AUDIENCE: Are you taking into account 00:10:23.342 --> 00:10:25.960 that that grain still has to pass radiological standards 00:10:25.960 --> 00:10:26.980 for the European-- 00:10:26.980 --> 00:10:28.547 R. SCOTT KEMP: It will. 00:10:28.547 --> 00:10:29.380 AUDIENCE: Say again. 00:10:29.380 --> 00:10:29.800 R. SCOTT KEMP: It will. 00:10:29.800 --> 00:10:30.300 Yeah. 00:10:30.300 --> 00:10:32.500 It still passes the standards. 00:10:32.500 --> 00:10:36.480 It's just that it's going to be higher than it was. 00:10:36.480 --> 00:10:38.380 So that's what we're looking at. 00:10:38.380 --> 00:10:40.680 The effect is very, very small amounts 00:10:40.680 --> 00:10:44.660 of dose times very large numbers of people. 00:10:44.660 --> 00:10:46.800 And that's what we have to deal with. 00:10:50.140 --> 00:10:53.900 Here's Fukushima. 00:10:53.900 --> 00:10:56.860 21 petabecquerels instead of 85. 00:10:56.860 --> 00:10:59.820 It's a BWR type reactor, water moderated, 00:10:59.820 --> 00:11:03.780 high-pressure containment, much more similar to the reactors 00:11:03.780 --> 00:11:05.780 that we have operating today. 00:11:05.780 --> 00:11:09.360 There are-- I have to double-check this. 00:11:09.360 --> 00:11:12.725 I think there are 94 of these reactors in the world, 00:11:12.725 --> 00:11:14.100 but I don't think all of them are 00:11:14.100 --> 00:11:17.980 operating because a lot of them were shut down. 00:11:17.980 --> 00:11:20.200 But it's higher than 11. 00:11:22.980 --> 00:11:25.900 The accident at Fukushima was really 00:11:25.900 --> 00:11:28.520 initiated by the tsunami that followed the earthquake. 00:11:28.520 --> 00:11:30.060 The reactors were OK. 00:11:30.060 --> 00:11:32.540 After the earthquake, one of the reactors had been-- 00:11:32.540 --> 00:11:34.748 they were already shut down because of the earthquake 00:11:34.748 --> 00:11:36.940 for about an hour. 00:11:36.940 --> 00:11:40.390 But then the cooling systems failed. 00:11:40.390 --> 00:11:45.310 And as a result, the fuel got hotter, and eventually 00:11:45.310 --> 00:11:47.750 got hot enough to boil away the water 00:11:47.750 --> 00:11:51.750 and begin to melt and release fission products. 00:11:51.750 --> 00:11:56.790 Although the Fukushima reactor did have a small containment, 00:11:56.790 --> 00:12:04.230 hydrogen caused explosions, and that eventually, caused exposure 00:12:04.230 --> 00:12:07.550 to the spent fuel pool, which we talked about could 00:12:07.550 --> 00:12:08.970 have caught on fire. 00:12:08.970 --> 00:12:11.950 And NRC says if it had, the situation 00:12:11.950 --> 00:12:15.470 would have been 100 times worse or 25 times worse 00:12:15.470 --> 00:12:16.510 than Chernobyl. 00:12:16.510 --> 00:12:18.470 Fortunately, that didn't happen. 00:12:18.470 --> 00:12:20.410 But they still had to vent the containment, 00:12:20.410 --> 00:12:26.250 the high-pressure containment, to prevent it from rupturing. 00:12:26.250 --> 00:12:28.430 And so that's where the releases-- 00:12:28.430 --> 00:12:33.150 significant releases came from these venting events. 00:12:33.150 --> 00:12:39.330 So the total amount released was still much smaller. 00:12:39.330 --> 00:12:47.150 And here is the situation for the releases from Fukushima. 00:12:47.150 --> 00:12:50.670 So you see these momentary pulses where they release some, 00:12:50.670 --> 00:12:52.170 and then they close the containment. 00:12:52.170 --> 00:12:57.610 Again, they release some more and so on. 00:12:57.610 --> 00:13:01.750 Let me see if I can speed this along. 00:13:01.750 --> 00:13:02.250 Yeah. 00:13:04.970 --> 00:13:07.330 So you can see, the lucky thing about Fukushima 00:13:07.330 --> 00:13:10.810 is the winds off the Kanto plane of Japan 00:13:10.810 --> 00:13:13.850 almost always blow into the Pacific Ocean. 00:13:13.850 --> 00:13:18.210 And so we really got lucky where these very high-dose plumes that 00:13:18.210 --> 00:13:21.250 might have contaminated a large land area, 00:13:21.250 --> 00:13:24.010 basically all just got washed into the ocean. 00:13:24.010 --> 00:13:27.050 However, a lot did get to the United States and Canada. 00:13:27.050 --> 00:13:31.570 And I'll show you-- here is the final deposited contamination. 00:13:31.570 --> 00:13:35.850 And it turns out that the United States probably got more-- 00:13:35.850 --> 00:13:37.310 I mean, the North America probably 00:13:37.310 --> 00:13:45.310 got more contamination in total than Japan from this accident. 00:13:45.310 --> 00:13:47.710 The thing I want you to take away from this chart, 00:13:47.710 --> 00:13:50.390 again, keeping in mind linear no threshold, 00:13:50.390 --> 00:13:53.790 is that the contamination here is 00:13:53.790 --> 00:13:56.770 about, let's say, 1% of the contamination, 00:13:56.770 --> 00:13:59.030 the average contamination here-- 00:13:59.030 --> 00:14:01.830 something of order, 1%. 00:14:01.830 --> 00:14:07.470 But the land area here is roughly of order 100 times 00:14:07.470 --> 00:14:10.990 the land area that is contaminated over here. 00:14:10.990 --> 00:14:13.670 So we could say, very roughly, since we're 00:14:13.670 --> 00:14:15.610 doing a back of the envelope calculation, 00:14:15.610 --> 00:14:19.470 that the global contamination might be like a factor of 2, 00:14:19.470 --> 00:14:22.210 higher than the local contamination. 00:14:22.210 --> 00:14:23.367 Something like that. 00:14:23.367 --> 00:14:24.950 It turns out you don't have to rely on 00:14:24.950 --> 00:14:29.070 of rough hand-waving argument. 00:14:29.070 --> 00:14:30.990 UNSCEAR has actually done a calculation 00:14:30.990 --> 00:14:32.310 I will show you later. 00:14:32.310 --> 00:14:34.920 And when they compared Europe to global, 00:14:34.920 --> 00:14:37.007 they found a factor of 1.6. 00:14:37.007 --> 00:14:38.340 It depends where the oceans are. 00:14:38.340 --> 00:14:39.420 Depends where the winds are blowing. 00:14:39.420 --> 00:14:41.040 Depends where the reactors are. 00:14:41.040 --> 00:14:42.820 All these things matter. 00:14:42.820 --> 00:14:46.360 But there's roughly, very roughly, a factor of 2-- 00:14:46.360 --> 00:14:48.320 less than a digit of significance. 00:14:48.320 --> 00:14:51.140 So that's a good thing to keep in your mind. 00:14:51.140 --> 00:14:57.400 So what is a, quote, "typical" major in INES 7 reactor 00:14:57.400 --> 00:14:57.900 accident? 00:14:57.900 --> 00:14:58.720 Yeah? 00:14:58.720 --> 00:15:00.095 AUDIENCE: I have a quick question 00:15:00.095 --> 00:15:04.120 about how much is 1 kilo becquerel 00:15:04.120 --> 00:15:07.140 per square meter compared to just background radiation? 00:15:12.280 --> 00:15:15.040 R. SCOTT KEMP: I think background contamination 00:15:15.040 --> 00:15:19.600 from nuclear weapons testing is about a third of a-- 00:15:19.600 --> 00:15:21.820 or half a kilo becquerel per square meter. 00:15:21.820 --> 00:15:22.900 I've could go look it up. 00:15:27.100 --> 00:15:28.700 If we have time, we'll do-- let me-- 00:15:28.700 --> 00:15:30.260 we'll do it at the end of class. 00:15:33.200 --> 00:15:39.380 So here's what I propose for estimating a typical reactor 00:15:39.380 --> 00:15:40.993 accident. 00:15:40.993 --> 00:15:42.660 There's still some Chernobyl-type things 00:15:42.660 --> 00:15:45.140 around that could cause bad accidents. 00:15:45.140 --> 00:15:48.120 We have 450 reactors, but 11 of these left. 00:15:48.120 --> 00:15:52.000 Let's just weight it-- a little bit of Chernobyl in there. 00:15:52.000 --> 00:15:54.060 Not a lot of Chernobyl, but a little bit. 00:15:54.060 --> 00:15:58.540 And then we'll have the Fukushima, accident B, 00:15:58.540 --> 00:16:01.300 which describes most reactors in the world. 00:16:01.300 --> 00:16:04.420 B is typical, but we'll divide that by 3. 00:16:04.420 --> 00:16:08.140 Because it's actually three reactor accidents-- 00:16:08.140 --> 00:16:12.140 one site accident, but three reactor accidents. 00:16:12.140 --> 00:16:15.360 So we have to make sure we do our accounting correctly. 00:16:15.360 --> 00:16:17.180 So we'll take that release, which 21, 00:16:17.180 --> 00:16:19.140 and we'll divide that by 3. 00:16:19.140 --> 00:16:21.740 So here's the calculation. 00:16:21.740 --> 00:16:23.540 85 becquerels times this weighting. 00:16:23.540 --> 00:16:25.820 21 becquerels times that weighting. 00:16:25.820 --> 00:16:29.340 And it's about 9 petabecquerels for what I'm going 00:16:29.340 --> 00:16:32.590 to call a typical accident? 00:16:32.590 --> 00:16:33.470 Yep? 00:16:33.470 --> 00:16:35.830 AUDIENCE: Where does that number 450 come from? 00:16:35.830 --> 00:16:36.530 What is it-- 00:16:36.530 --> 00:16:38.990 R. SCOTT KEMP: Oh, total number of reactors in the world. 00:16:38.990 --> 00:16:41.630 Yep. 00:16:41.630 --> 00:16:44.450 This is just my weighting. 00:16:44.450 --> 00:16:46.390 So this is per reactor accident. 00:16:46.390 --> 00:16:53.230 Now, it turns out that not all accidents are simple reactor 00:16:53.230 --> 00:16:54.350 accidents. 00:16:54.350 --> 00:16:59.150 Some accidents are internal accidents, like at Chernobyl, 00:16:59.150 --> 00:17:01.070 and some accidents are site accidents 00:17:01.070 --> 00:17:03.270 that involve multiple reactors. 00:17:03.270 --> 00:17:07.510 So we need to figure out how to weight them. 00:17:07.510 --> 00:17:09.010 So what I'm going to do is just say, 00:17:09.010 --> 00:17:11.710 well, let's just use empirical data. 00:17:11.710 --> 00:17:16.109 Barring any better data, roughly half of them will be internal, 00:17:16.109 --> 00:17:18.724 and roughly half of them will be external. 00:17:18.724 --> 00:17:20.349 And actually, that makes a lot of sense 00:17:20.349 --> 00:17:23.310 from a kind of statistics argument, in the sense 00:17:23.310 --> 00:17:27.670 that if it turned out internal accidents dominated 00:17:27.670 --> 00:17:32.490 tremendously, and the types of accidents of external events 00:17:32.490 --> 00:17:36.070 did not, that means you haven't done your safety well. 00:17:36.070 --> 00:17:37.890 It means you've underperformed in one area 00:17:37.890 --> 00:17:39.770 and overdone it in another area. 00:17:39.770 --> 00:17:43.850 So we expect these things to be roughly the same 00:17:43.850 --> 00:17:45.810 if we design our safety. 00:17:45.810 --> 00:17:47.790 So here's my calculation-- 00:17:47.790 --> 00:17:50.630 1/2 internal plus 1/2 site accident. 00:17:50.630 --> 00:17:52.250 So I'm going to put the 3 back in. 00:17:52.250 --> 00:17:53.970 I get 18 petabecquerel. 00:17:53.970 --> 00:17:59.090 And it's about 1/5 of our Chernobyl event, very roughly. 00:17:59.090 --> 00:18:02.770 That's what I'm going to argue is a typical accident. 00:18:02.770 --> 00:18:04.570 You don't have to agree with this. 00:18:04.570 --> 00:18:08.810 Do whatever you think is right and propagate your own math. 00:18:08.810 --> 00:18:13.370 But this is just like-- this is what the empirical history says. 00:18:13.370 --> 00:18:19.370 So once we have that accident, then what happens? 00:18:19.370 --> 00:18:22.530 Well, there are lots of different pathways 00:18:22.530 --> 00:18:25.050 in which you can be irradiated. 00:18:25.050 --> 00:18:30.670 You can have stuff landing on the ground. 00:18:30.670 --> 00:18:34.070 And then it just sits there and it glows and it irradiates you. 00:18:34.070 --> 00:18:36.070 Call this ground shine. 00:18:36.070 --> 00:18:40.950 You could inhale it because it's floating around in the air. 00:18:40.950 --> 00:18:42.950 And that inhalation can come directly 00:18:42.950 --> 00:18:45.670 because it's in the air after the accident, 00:18:45.670 --> 00:18:47.030 or you kick up some dust. 00:18:47.030 --> 00:18:50.230 Roads and stuff is constantly kicking up dust, 00:18:50.230 --> 00:18:51.670 and you Inhale it. 00:18:51.670 --> 00:18:56.670 Or it could go into the food supply, and you eat it. 00:18:56.670 --> 00:19:00.150 So we have three different pathways here. 00:19:00.150 --> 00:19:03.430 And it turns out that these pathways 00:19:03.430 --> 00:19:08.190 are complicated by the chemistry of the isotope that 00:19:08.190 --> 00:19:09.590 is being released. 00:19:09.590 --> 00:19:12.630 Once it goes into the environment, how 00:19:12.630 --> 00:19:16.670 that isotope operates depends on all the chemical reactions 00:19:16.670 --> 00:19:18.390 that are happening out there. 00:19:18.390 --> 00:19:20.870 So as it turns out, this is not something you 00:19:20.870 --> 00:19:22.430 can do from first principles. 00:19:22.430 --> 00:19:29.800 You need to have basically experimental data that tells us, 00:19:29.800 --> 00:19:32.600 on average, what does the chemistry of the environment 00:19:32.600 --> 00:19:36.200 do to the availability of these isotopes? 00:19:36.200 --> 00:19:39.060 So that data has been done. 00:19:39.060 --> 00:19:44.640 This is from a 1982 report by the United Nations. 00:19:44.640 --> 00:19:53.720 And they calculate that for every becquerel meter squared 00:19:53.720 --> 00:20:08.760 of contamination, there's 0.89, or 9 nanograys of contamination. 00:20:08.760 --> 00:20:12.480 We want a kilobecquerel. 00:20:12.480 --> 00:20:19.520 So we actually could state this as 9 micrograys 00:20:19.520 --> 00:20:21.700 per kilobecquerel per meter squared. 00:20:21.700 --> 00:20:25.920 So we just go back to our situation, 00:20:25.920 --> 00:20:32.300 we have the original contamination 00:20:32.300 --> 00:20:36.460 is 20% of Chernobyl. 00:20:36.460 --> 00:20:40.600 The Chernobyl was 10 kilobecquerel per square meter. 00:20:46.610 --> 00:20:50.040 And now we have this conversion factor, 00:20:50.040 --> 00:20:54.180 which is 9 microsieverts, or 9 microgray 00:20:54.180 --> 00:20:55.820 is what it actually says-- we'll talk 00:20:55.820 --> 00:21:01.940 about that in a second-- per kilobecquerel per meter squared. 00:21:01.940 --> 00:21:04.860 Does that sound right? 00:21:04.860 --> 00:21:08.220 And then this is all decaying. 00:21:08.220 --> 00:21:12.300 And it's only cesium-137, so we'll 00:21:12.300 --> 00:21:16.660 have some time dependency, which is the half life. 00:21:16.660 --> 00:21:21.900 And then this is 30.2 years because this 00:21:21.900 --> 00:21:24.600 is the half life of cesium-137. 00:21:24.600 --> 00:21:32.240 So this is the dose as a function of time. 00:21:32.240 --> 00:21:34.960 Does that make sense? 00:21:34.960 --> 00:21:36.460 From a typical accident? 00:21:57.540 --> 00:21:58.040 Yes? 00:22:01.487 --> 00:22:03.820 Said a typical accident is about 1/5 fifth of Chernobyl. 00:22:03.820 --> 00:22:05.778 Chernobyl is 10 kilobecquerels per square meter 00:22:05.778 --> 00:22:07.720 of contamination for cesium. 00:22:07.720 --> 00:22:12.480 And the dose is 9 microsieverts per kilobecquerel per-- 00:22:12.480 --> 00:22:14.140 9 microsieverts per year. 00:22:14.140 --> 00:22:14.760 I'm sorry. 00:22:14.760 --> 00:22:22.570 9 microsieverts per year per kilobecquerel per meter squared. 00:22:22.570 --> 00:22:29.090 And then it decays, and so there's a decay factor 00:22:29.090 --> 00:22:35.370 It turns out this number is really important. 00:22:35.370 --> 00:22:39.190 It affects the result a lot. 00:22:39.190 --> 00:22:41.650 And it is a number that is not fully agreed 00:22:41.650 --> 00:22:45.050 upon in the literature. 00:22:45.050 --> 00:22:49.050 I'll just briefly mention, we-- 00:22:49.050 --> 00:22:53.130 I just said there's a one to one conversion between grays 00:22:53.130 --> 00:22:56.450 and sieverts, which we know is the relative biological 00:22:56.450 --> 00:23:00.710 effectiveness for beta radiation. 00:23:00.710 --> 00:23:03.690 And gammas are about the same, one to one. 00:23:03.690 --> 00:23:09.010 Technically, beta radiation is slightly higher. 00:23:09.010 --> 00:23:12.930 But if we remember the decay from cesium, 00:23:12.930 --> 00:23:15.610 most of the beta energy is carried away 00:23:15.610 --> 00:23:17.250 by the antineutrino. 00:23:17.250 --> 00:23:20.390 So for the most part, this is a gamma dose. 00:23:20.390 --> 00:23:23.650 So we really can just call sieverts and grays the same. 00:23:27.210 --> 00:23:32.810 So 9 microsieverts is basically what this says. 00:23:32.810 --> 00:23:37.030 Now, that's only the ground shine dose. 00:23:37.030 --> 00:23:40.370 This is dose commitments due to-- see the top of the slide-- 00:23:40.370 --> 00:23:42.590 external irradiation. 00:23:42.590 --> 00:23:46.470 What about the stuff that you eat or that you breathe? 00:23:46.470 --> 00:23:49.610 Well, the UN also has data for that. 00:23:49.610 --> 00:23:53.390 They have these external radiation inhalation ingestion 00:23:53.390 --> 00:23:54.670 in total. 00:23:54.670 --> 00:23:57.830 And so what we can do is we can just 00:23:57.830 --> 00:24:01.990 take-- we can calculate the other two by just taking 00:24:01.990 --> 00:24:06.190 the total, which is given as 220, 00:24:06.190 --> 00:24:12.430 and dividing it by the external, which is 150. 00:24:12.430 --> 00:24:14.870 And that is basically a weighting factor 00:24:14.870 --> 00:24:30.640 that converts the external only to the total dose 00:24:30.640 --> 00:24:33.920 from the other ingestion pathways. 00:24:33.920 --> 00:24:37.440 People look confused. 00:24:37.440 --> 00:24:41.520 Who is slightly confused? 00:24:41.520 --> 00:24:42.060 You are. 00:24:42.060 --> 00:24:42.580 OK. 00:24:42.580 --> 00:24:47.760 So what we did is we-- the first thing we did is we calculated-- 00:24:47.760 --> 00:24:50.240 are we good up to here? 00:24:50.240 --> 00:24:52.320 OK, we're good up to the dose coming off 00:24:52.320 --> 00:24:55.580 the ground per unit time. 00:24:55.580 --> 00:24:56.620 And then it decays. 00:24:56.620 --> 00:24:58.000 We're good with that. 00:24:58.000 --> 00:25:01.040 So now, this is just the dose coming off the ground. 00:25:01.040 --> 00:25:06.560 And they have this chart that says for 150 times 10 00:25:06.560 --> 00:25:11.040 to the 4 man sieverts from external irradiation, 00:25:11.040 --> 00:25:14.040 we would have 220 total. 00:25:14.040 --> 00:25:17.460 And basically, this is just a scaling factor. 00:25:17.460 --> 00:25:23.180 For every 150 from external, we would have 00:25:23.180 --> 00:25:30.060 150 external plus 70 internal. 00:25:30.060 --> 00:25:33.300 150 plus 70 is 220. 00:25:33.300 --> 00:25:35.300 So it's just a scaling factor to go up 00:25:35.300 --> 00:25:40.555 to include the effect of the internal dose. 00:25:40.555 --> 00:25:41.180 And so I just-- 00:25:41.180 --> 00:25:42.860 220 over 150. 00:25:42.860 --> 00:25:46.860 I think I hope that gets you where we need to be. 00:25:46.860 --> 00:25:49.700 AUDIENCE: But if it's external dose, and that 00:25:49.700 --> 00:25:52.860 comes directly from Chernobyl, then 00:25:52.860 --> 00:25:56.500 wouldn't the amount that you ingest be less, not more? 00:25:56.500 --> 00:25:58.100 R. SCOTT KEMP: It is much less. 00:25:58.100 --> 00:26:03.340 It's about 69. 00:26:03.340 --> 00:26:11.740 But when you add 69 plus 150, you get 220. 00:26:11.740 --> 00:26:14.333 AUDIENCE: Because that number is only external. 00:26:14.333 --> 00:26:16.000 R. SCOTT KEMP: That number we calculated 00:26:16.000 --> 00:26:17.740 was only the ground shine dose. 00:26:17.740 --> 00:26:19.680 That's what this plot said here. 00:26:19.680 --> 00:26:22.240 This is external irradiation. 00:26:22.240 --> 00:26:23.220 You see at the top? 00:26:23.220 --> 00:26:24.640 AUDIENCE: [INAUDIBLE]. 00:26:24.640 --> 00:26:26.140 But kilometers there. 00:26:26.140 --> 00:26:28.800 But-- sorry-- kilobecquerels there. 00:26:28.800 --> 00:26:31.120 R. SCOTT KEMP: Yeah, so I converted-- 00:26:31.120 --> 00:26:44.120 so this would be 9.9 times 10 to the 8, which is 9 nanogray. 00:26:44.120 --> 00:26:50.400 And 9 nanogray per becquerel is the same as 9 00:26:50.400 --> 00:26:52.807 microgray per kilobecquerel. 00:26:52.807 --> 00:26:53.640 AUDIENCE: All right. 00:26:53.640 --> 00:26:54.640 Thanks. 00:26:54.640 --> 00:26:57.200 R. SCOTT KEMP: Yep. 00:26:57.200 --> 00:26:59.960 Unit conversion. 00:26:59.960 --> 00:27:04.800 So this is our formula. 00:27:04.800 --> 00:27:08.480 And now we just need to say how many people were exposed. 00:27:08.480 --> 00:27:11.320 So it turns out the population of Europe 00:27:11.320 --> 00:27:14.610 has been very stable for a long time-- 00:27:14.610 --> 00:27:18.810 730 to 750 million people. 00:27:18.810 --> 00:27:21.530 We can basically take it as a fixed value for the sake 00:27:21.530 --> 00:27:23.010 of our calculation. 00:27:23.010 --> 00:27:28.130 If you wanted to-- it's stable over the lifetime of what? 00:27:28.130 --> 00:27:31.350 Over several of these, several half lives. 00:27:31.350 --> 00:27:32.490 So 100 years. 00:27:32.490 --> 00:27:34.130 That's reasonable. 00:27:34.130 --> 00:27:36.010 If we were doing this in Africa, we 00:27:36.010 --> 00:27:38.497 would need to have a time-dependent population, 00:27:38.497 --> 00:27:40.330 some place where the population dynamics are 00:27:40.330 --> 00:27:41.750 changing dramatically. 00:27:41.750 --> 00:27:44.290 But we just treat it as stable. 00:27:44.290 --> 00:27:48.470 So 750 times 10 to the 6 people. 00:27:52.970 --> 00:27:59.530 And then we're going to integrate that from 0 00:27:59.530 --> 00:28:01.010 to infinity. 00:28:01.010 --> 00:28:03.050 Now, going to infinity is not a big deal 00:28:03.050 --> 00:28:06.110 because almost all of it is decayed after some half life. 00:28:06.110 --> 00:28:10.010 It just adds little tiny bits at the end. 00:28:10.010 --> 00:28:12.170 So let's just cancel our units. 00:28:12.170 --> 00:28:14.070 Kilobecquerels cancels. 00:28:14.070 --> 00:28:15.530 Meter squares cancel. 00:28:19.110 --> 00:28:23.110 These years will cancel with these years. 00:28:23.110 --> 00:28:26.710 And what we'll be left with is something 00:28:26.710 --> 00:28:45.370 that has units of sieverts times persons, something like that, 00:28:45.370 --> 00:28:48.030 what we call person sieverts. 00:28:48.030 --> 00:28:50.710 And so if you do the math, and you do the integral, 00:28:50.710 --> 00:28:56.030 you get 860,000 person sieverts of exposure. 00:28:56.030 --> 00:28:58.190 OK? 00:28:58.190 --> 00:28:58.910 Yep? 00:28:58.910 --> 00:28:59.577 AUDIENCE: Sorry. 00:28:59.577 --> 00:29:01.833 What does person sievert mean? 00:29:01.833 --> 00:29:03.250 R. SCOTT KEMP: Remember, because-- 00:29:03.250 --> 00:29:07.430 AUDIENCE: Sievert per person [INAUDIBLE]. 00:29:07.430 --> 00:29:09.280 R. SCOTT KEMP: Sieverts per person-- 00:29:09.280 --> 00:29:11.812 it's sieverts of dose times the number of people. 00:29:11.812 --> 00:29:13.520 So remember, because we're in the linear, 00:29:13.520 --> 00:29:15.440 we can use linear no threshold model. 00:29:21.760 --> 00:29:25.520 Say, a millisievert for 10 people 00:29:25.520 --> 00:29:29.280 would be 10 milli person sieverts. 00:29:34.600 --> 00:29:37.960 AUDIENCE: If I have 10 millisieverts or whatever dose, 00:29:37.960 --> 00:29:41.960 and I have 10 people, I expect that that 10 millisievert 00:29:41.960 --> 00:29:43.930 is divided by the number. 00:29:43.930 --> 00:29:45.180 R. SCOTT KEMP: No, no, no, no. 00:29:45.180 --> 00:29:47.160 AUDIENCE: [INAUDIBLE] 00:29:47.160 --> 00:29:51.272 R. SCOTT KEMP: No, because we're not taking the radiation-- 00:29:51.272 --> 00:29:53.980 That would be the case if I were taking the radiation and saying, 00:29:53.980 --> 00:29:55.680 OK, you have a little bit, and you can have a little bit, 00:29:55.680 --> 00:29:57.160 and you can have a little bit. 00:29:57.160 --> 00:29:58.960 But that's not what's happening here. 00:29:58.960 --> 00:30:00.960 What's happening is I'm spreading the radiation 00:30:00.960 --> 00:30:02.280 everywhere. 00:30:02.280 --> 00:30:04.760 We have a background of radiation. 00:30:04.760 --> 00:30:08.960 And if I put one person there, I get one person sievert of dose. 00:30:08.960 --> 00:30:12.060 And if I put two people in that same field of radiation, 00:30:12.060 --> 00:30:15.193 unless they're standing so close that their self-shielding. 00:30:15.193 --> 00:30:17.860 But as long as I have one person here, and one person over here, 00:30:17.860 --> 00:30:20.980 and one person over here, the more people I add, 00:30:20.980 --> 00:30:25.440 the more person sieverts of dose I'm creating. 00:30:25.440 --> 00:30:29.360 And so I give this person this 0.1% chance of cancer, 00:30:29.360 --> 00:30:31.120 and this person 0.1% chance of cancer, 00:30:31.120 --> 00:30:33.200 and this 1.1% chance of cancer. 00:30:33.200 --> 00:30:35.340 And now it's total cancers are going 00:30:35.340 --> 00:30:41.000 to be 0.1 plus 0.1 plus 0.1 total cancers. 00:30:41.000 --> 00:30:43.700 AUDIENCE: So if I take [INAUDIBLE] millisievert, 00:30:43.700 --> 00:30:45.820 then the person next to me [INAUDIBLE]. 00:30:45.820 --> 00:30:47.700 R. SCOTT KEMP: Yes. 00:30:47.700 --> 00:30:48.740 It doesn't affect-- 00:30:48.740 --> 00:30:50.380 AUDIENCE: The more I take, the less [INAUDIBLE]. 00:30:50.380 --> 00:30:51.140 R. SCOTT KEMP: That's not-- yeah, 00:30:51.140 --> 00:30:52.360 the first description is correct. 00:30:52.360 --> 00:30:53.610 The second one is not correct. 00:30:53.610 --> 00:30:55.260 Yeah, exactly. 00:30:55.260 --> 00:30:58.740 Because it's just spread environmentally, 00:30:58.740 --> 00:31:01.440 not handed out like pieces of cake. 00:31:04.740 --> 00:31:06.020 It is not yellowcake. 00:31:09.200 --> 00:31:14.760 So everyone good with this unit, person sievert? 00:31:14.760 --> 00:31:17.400 Now we need to convert this to cancers. 00:31:17.400 --> 00:31:19.060 So we actually know how to do this. 00:31:19.060 --> 00:31:24.200 We looked at this in class already. 00:31:24.200 --> 00:31:27.680 We can say-- if we have-- 00:31:27.680 --> 00:31:28.700 I'll go over here. 00:31:33.320 --> 00:31:41.400 We have 860,000 person sieverts. 00:31:41.400 --> 00:31:46.800 And we saw that the excess relative risk of cancer 00:31:46.800 --> 00:31:52.800 was 0.64 per sievert. 00:31:52.800 --> 00:31:57.640 So the excess relative risk-- 00:31:57.640 --> 00:31:58.140 sorry. 00:31:58.140 --> 00:32:03.670 The excess risk is the baseline risk. 00:32:03.670 --> 00:32:07.770 Does anyone remember what the baseline risk is for cancer? 00:32:07.770 --> 00:32:09.130 AUDIENCE: 20%. 00:32:09.130 --> 00:32:17.610 R. SCOTT KEMP: 20% times 0.64 per sievert. 00:32:17.610 --> 00:32:25.250 And so if we take this, and we multiply it by 0.64 per sievert 00:32:25.250 --> 00:32:34.850 times 0.0.2 for 20%, this cancels. 00:32:34.850 --> 00:32:38.090 This is a m and this is cancers. 00:32:38.090 --> 00:32:43.690 20% is cancers per person. 00:32:43.690 --> 00:32:45.530 That's the unit for the baseline risk. 00:32:50.450 --> 00:32:54.250 And so we have-- 00:32:54.250 --> 00:32:58.890 this number-- let's see-- person's cancer, 00:32:58.890 --> 00:33:00.710 and sieverts are canceled. 00:33:00.710 --> 00:33:03.710 So we should be able to get the total number of cancers, 00:33:03.710 --> 00:33:14.270 which is 110,000 cancers. 00:33:14.270 --> 00:33:17.150 And you may not know this, but I know this. 00:33:17.150 --> 00:33:19.790 About half of cancers are fatal. 00:33:19.790 --> 00:33:28.190 So this is 55,000 deaths-- 00:33:28.190 --> 00:33:36.970 Europe only, cesium-137 only. 00:33:46.870 --> 00:33:50.770 We don't have to take these people's word for it. 00:33:50.770 --> 00:33:53.270 We can also go to the National Academies. 00:33:53.270 --> 00:33:56.010 They have this fancy supermodel. 00:33:56.010 --> 00:33:57.150 Yes? 00:33:57.150 --> 00:34:00.163 AUDIENCE: This is per--like this accident. 00:34:00.163 --> 00:34:01.330 R. SCOTT KEMP: Per accident. 00:34:01.330 --> 00:34:03.530 Per typical accident. 00:34:03.530 --> 00:34:05.350 1/5 fifth of a Chernobyl. 00:34:05.350 --> 00:34:05.850 Yes. 00:34:09.210 --> 00:34:11.030 So you don't have to use that plot 00:34:11.030 --> 00:34:13.489 and do it the way I just did it from first principles. 00:34:13.489 --> 00:34:15.290 You could look up one of these tables. 00:34:15.290 --> 00:34:18.610 They're here on page 280 of the National Academy's committee 00:34:18.610 --> 00:34:21.130 to assess the health risk from exposure 00:34:21.130 --> 00:34:24.010 to load levels of ionizing radiation, which 00:34:24.010 --> 00:34:27.090 is called BEIR VII. 00:34:27.090 --> 00:34:30.730 They have this big 100, multipage, huge document 00:34:30.730 --> 00:34:33.210 where they try to calculate all the different risk models. 00:34:33.210 --> 00:34:34.830 And they have a relative risk model, 00:34:34.830 --> 00:34:37.250 and they have an absolute risk transport model. 00:34:37.250 --> 00:34:39.093 And they have it sex organized. 00:34:39.093 --> 00:34:41.010 And they have it for all the different organs, 00:34:41.010 --> 00:34:42.750 and all these different things. 00:34:42.750 --> 00:34:45.929 And they can add it all up, and they give you these numbers. 00:34:45.929 --> 00:34:48.150 At the end of the day, for all solid cancers, 00:34:48.150 --> 00:34:56.929 480 per 100,000 people exposed to 0.1 gray. 00:34:56.929 --> 00:35:01.980 So that's 480 solid cancers in-- this is for men-- 00:35:01.980 --> 00:35:05.860 per 10,000 person gray. 00:35:05.860 --> 00:35:07.540 And here, we have the same thing. 00:35:07.540 --> 00:35:09.240 704-- sorry, ladies. 00:35:09.240 --> 00:35:13.300 You're much more likely to die. 00:35:13.300 --> 00:35:14.720 This also is not cancers. 00:35:14.720 --> 00:35:16.080 This goes directly to mortality. 00:35:16.080 --> 00:35:17.260 You see at the top-- 00:35:17.260 --> 00:35:18.480 solid cancer mortality. 00:35:18.480 --> 00:35:21.740 So this is their statistical data on just mortality. 00:35:21.740 --> 00:35:27.980 740 per 10,000 person gray. 00:35:27.980 --> 00:35:29.300 And if we add averages. 00:35:29.300 --> 00:35:32.620 We say roughly, it's 50% men and women. 00:35:32.620 --> 00:35:36.580 Then 610 is the average of these two numbers. 00:35:36.580 --> 00:35:42.740 610 deaths per 100,000 person gray. 00:35:42.740 --> 00:35:44.600 So let's do that here. 00:35:52.500 --> 00:36:01.520 610 deaths per 100,000 person gray. 00:36:01.520 --> 00:36:04.240 And how many person gray did we have? 00:36:04.240 --> 00:36:10.480 860,000 person gray, which is 86 times 110,000. 00:36:10.480 --> 00:36:13.126 So times 86-- 00:36:13.126 --> 00:36:17.520 860,000 person grays. 00:36:17.520 --> 00:36:26.560 And 610 times 86 is, I think, 520 deaths. 00:36:26.560 --> 00:36:30.360 So it basically matches that calculation. 00:36:30.360 --> 00:36:33.720 That's nice to know that they both match. 00:36:33.720 --> 00:36:38.560 So pretty close. 00:36:38.560 --> 00:36:44.440 So this is only cesium. 00:36:44.440 --> 00:36:45.860 Was this what? 00:36:45.860 --> 00:36:47.080 AUDIENCE: [INAUDIBLE] 00:36:47.080 --> 00:36:48.540 R. SCOTT KEMP: Oh, no. 00:36:52.920 --> 00:36:54.340 That is wrong by-- 00:36:56.970 --> 00:37:01.850 that is wrong by my math. 00:37:01.850 --> 00:37:07.550 What is 600-- they're not off by a factor of 10. 00:37:07.550 --> 00:37:08.510 What am I doing wrong? 00:37:08.510 --> 00:37:12.906 AUDIENCE: A hundred [? thousands. ?] 00:37:23.810 --> 00:37:25.310 R. SCOTT KEMP: Who has a calculator? 00:37:27.930 --> 00:37:30.210 86 times 610. 00:37:30.210 --> 00:37:33.108 AUDIENCE: [INAUDIBLE] 00:37:35.657 --> 00:37:36.490 R. SCOTT KEMP: Typo. 00:37:39.570 --> 00:37:41.610 Yeah, they match. 00:37:41.610 --> 00:37:43.790 If they didn't match, I wouldn't have presented it. 00:37:47.090 --> 00:37:49.270 52,000 deaths, not 500,000 deaths. 00:37:49.270 --> 00:37:52.190 52,000 deaths. 00:37:52.190 --> 00:37:54.650 So that was only-- this is only from cesium. 00:37:54.650 --> 00:37:58.510 This is only in Europe from a typical accident. 00:37:58.510 --> 00:38:00.430 So I mentioned that the other thing 00:38:00.430 --> 00:38:02.590 that we have to really worry about is iodine. 00:38:02.590 --> 00:38:05.790 And so let's just look at what the situation was iodine. 00:38:05.790 --> 00:38:07.910 This is from an UNSCEAR report. 00:38:07.910 --> 00:38:10.670 And what they do is they actually-- 00:38:10.670 --> 00:38:13.970 they carefully calculate all the iodine exposure out there. 00:38:13.970 --> 00:38:16.630 There's a lot of experimental assessment. 00:38:16.630 --> 00:38:20.930 And they have categories of evacuees, recovery workers, 00:38:20.930 --> 00:38:24.630 inhabitants of Belarus, Russia, and Ukraine, 00:38:24.630 --> 00:38:26.610 and inhabitants of distant countries. 00:38:26.610 --> 00:38:29.350 And they figure it all out what the dosing is. 00:38:29.350 --> 00:38:32.650 And they have the collective thyroid dose only in 1986, 00:38:32.650 --> 00:38:35.830 because it doesn't really last that long. 00:38:35.830 --> 00:38:37.810 And they find if you add all these up, 00:38:37.810 --> 00:38:42.070 it's about 3 million person sieverts. 00:38:42.070 --> 00:38:47.830 And they have also the effective dose, not to the thyroid, 00:38:47.830 --> 00:38:49.750 to other parts of your body. 00:38:49.750 --> 00:38:53.170 And it's about 380,000 person sieverts. 00:38:53.170 --> 00:38:58.490 So this is about 3.4, Something Like 3.4 million person 00:38:58.490 --> 00:38:59.933 sieverts of dose from iodine. 00:38:59.933 --> 00:39:01.850 And we don't have to do this whole decay thing 00:39:01.850 --> 00:39:03.310 because they've done all that. 00:39:03.310 --> 00:39:05.330 This thing is decayed all the way, 00:39:05.330 --> 00:39:07.810 and they're just giving you the final dose. 00:39:07.810 --> 00:39:09.290 So here it is. 00:39:09.290 --> 00:39:14.930 Here's the dose from Chernobyl. 00:39:14.930 --> 00:39:16.510 Here's the dose from cesium. 00:39:16.510 --> 00:39:17.970 Here's the dose from iodine. 00:39:17.970 --> 00:39:20.230 Here's some other isotopes that we could look at. 00:39:20.230 --> 00:39:23.090 And if I just rescale them all to our typical accident, 00:39:23.090 --> 00:39:24.510 here they are, rescaled. 00:39:27.890 --> 00:39:29.410 So I can add them all up together. 00:39:29.410 --> 00:39:30.910 There are a bunch of other isotopes, 00:39:30.910 --> 00:39:32.097 and they get really tiny. 00:39:32.097 --> 00:39:33.430 So we don't have to worry about. 00:39:33.430 --> 00:39:35.770 So let's just take the top four. 00:39:35.770 --> 00:39:39.090 And it turns out 2.3 million person sieverts is the total 00:39:39.090 --> 00:39:44.010 dose instead of 860,000, which means it's about 2.7 times 00:39:44.010 --> 00:39:50.100 higher than we calculated for total deaths. 00:39:50.100 --> 00:39:51.932 Yes? 00:39:51.932 --> 00:39:53.640 Everyone understand what's going on here? 00:39:53.640 --> 00:39:56.380 So we're just basically going from one isotope, 00:39:56.380 --> 00:39:58.620 and then I've done the same stuff 00:39:58.620 --> 00:40:00.900 for the other isotopes for you. 00:40:00.900 --> 00:40:05.220 And we can just go to the all isotope dose. 00:40:05.220 --> 00:40:11.940 And then UNSCEAR estimates that for Chernobyl, they 00:40:11.940 --> 00:40:14.780 estimated 10 million person sieverts of dose 00:40:14.780 --> 00:40:17.532 to all Europeans, 16 million person sieverts 00:40:17.532 --> 00:40:18.240 to all the world. 00:40:18.240 --> 00:40:21.500 So their multiplier to go from Europe to the rest of the world 00:40:21.500 --> 00:40:23.340 was a factor of 1.6. 00:40:23.340 --> 00:40:26.220 I earlier made the argument using 00:40:26.220 --> 00:40:30.300 that Fukushima map that it was roughly a factor of 2. 00:40:30.300 --> 00:40:32.900 This is a little bit more conservative, 00:40:32.900 --> 00:40:35.420 so let's use this number. 00:40:35.420 --> 00:40:40.220 So we would take our calculation of 55,000, or 52,000, deaths, 00:40:40.220 --> 00:40:44.540 and we would multiply it by 2.7, and multiply it by 1.6. 00:40:44.540 --> 00:40:49.640 So I average those together and suggested we use 53,000, 00:40:49.640 --> 00:40:56.960 and times 2.7 times 1.6, which gives us 230,000 deaths per 00:40:56.960 --> 00:41:02.570 typical major accident worldwide, all isotopes. 00:41:02.570 --> 00:41:03.560 Yes? 00:41:03.560 --> 00:41:04.680 AUDIENCE: And then what-- are we going to multiply that 00:41:04.680 --> 00:41:06.082 by the frequency of accidents? 00:41:06.082 --> 00:41:07.040 R. SCOTT KEMP: Correct. 00:41:07.040 --> 00:41:07.920 Yeah. 00:41:07.920 --> 00:41:12.320 So this is the basic result that you should have in your mind. 00:41:12.320 --> 00:41:17.240 The average large INES 7-- 00:41:17.240 --> 00:41:18.720 average according to this weighting 00:41:18.720 --> 00:41:20.320 that I showed you earlier-- 00:41:20.320 --> 00:41:24.040 accident kills a quarter million people from cancer 00:41:24.040 --> 00:41:26.120 and gives another quarter million people 00:41:26.120 --> 00:41:29.120 cancer that doesn't kill them. 00:41:29.120 --> 00:41:30.780 And that's only the cancer effects. 00:41:30.780 --> 00:41:34.160 This is not the teratogenic effects and inheritable defects, 00:41:34.160 --> 00:41:39.800 the deterministic diseases, all the other stuff, heart disease, 00:41:39.800 --> 00:41:41.300 et cetera. 00:41:41.300 --> 00:41:43.780 All the other all-hazard mortality is not included. 00:41:43.780 --> 00:41:46.250 This is only cancer deaths. 00:41:46.250 --> 00:41:48.910 But it's a big source of death. 00:41:48.910 --> 00:41:53.330 So it's a good number to go on. 00:41:53.330 --> 00:41:57.570 So quarter million people-- now, that sounds like a lot. 00:41:57.570 --> 00:41:59.330 That sounds like a lot. 00:41:59.330 --> 00:42:02.570 But you have to remember that coal is killing people all 00:42:02.570 --> 00:42:07.610 the time, and you just don't see these deaths from PM10, 00:42:07.610 --> 00:42:10.770 to PM2.5, from radiation from coal, et cetera. 00:42:10.770 --> 00:42:13.370 So we really don't know yet whether this 00:42:13.370 --> 00:42:15.390 is a big-- it's a very dramatic number, 00:42:15.390 --> 00:42:19.250 but we don't know if on average is big yet. 00:42:19.250 --> 00:42:21.710 AUDIENCE: So what is the time frame for these deaths? 00:42:21.710 --> 00:42:26.370 Because it's not clear to me whether this 00:42:26.370 --> 00:42:32.553 means there'll be a quarter of a million deaths in month after-- 00:42:32.553 --> 00:42:33.470 R. SCOTT KEMP: No, no. 00:42:33.470 --> 00:42:34.870 They are spread out. 00:42:34.870 --> 00:42:35.810 AUDIENCE: --20 years. 00:42:35.810 --> 00:42:38.730 R. SCOTT KEMP: They're spread out over 100 years. 00:42:38.730 --> 00:42:39.450 Yeah. 00:42:39.450 --> 00:42:44.670 They just appear as people who just got cancer. 00:42:44.670 --> 00:42:47.150 This is your aunt who got cancer. 00:42:47.150 --> 00:42:49.790 Why did they get cancer? 00:42:49.790 --> 00:42:53.590 And for some small fraction of the population, the reason 00:42:53.590 --> 00:42:59.950 they got cancer was radiation from cesium-137 00:42:59.950 --> 00:43:02.990 that just happened to shoot a photon into their body 00:43:02.990 --> 00:43:05.790 and break the genome in the wrong way. 00:43:05.790 --> 00:43:07.070 And you can't tell. 00:43:07.070 --> 00:43:10.950 There's no way that you can figure it out. 00:43:10.950 --> 00:43:14.110 And then because the background rate of cancers 00:43:14.110 --> 00:43:18.030 is already 1 in 5 persons, this is really 00:43:18.030 --> 00:43:22.030 just a tiny perturbation on that background. 00:43:22.030 --> 00:43:24.410 If you have 10 billion people, and 1/5 of them, 00:43:24.410 --> 00:43:27.390 2 billion people, are going to get cancer, 00:43:27.390 --> 00:43:29.630 this is a tiny perturbation. 00:43:29.630 --> 00:43:31.950 You can't see it. 00:43:31.950 --> 00:43:35.110 So that's the problem, is that you can't empirically 00:43:35.110 --> 00:43:37.870 see these deaths, but they're there. 00:43:37.870 --> 00:43:40.770 They are caused by the accident. 00:43:40.770 --> 00:43:41.270 Yeah? 00:43:41.270 --> 00:43:43.103 AUDIENCE: So how do we measure the, I guess, 00:43:43.103 --> 00:43:45.850 the frequency if the events are so sparse? 00:43:45.850 --> 00:43:47.032 [INAUDIBLE] 00:43:47.032 --> 00:43:48.490 R. SCOTT KEMP: Well, that goes back 00:43:48.490 --> 00:43:50.690 to how did we get the dose response model? 00:43:50.690 --> 00:43:55.770 And that is basically this study, the atomic bomb survivors 00:43:55.770 --> 00:43:57.330 plus radiation workers, people who 00:43:57.330 --> 00:44:01.290 have higher doses, where we then build a model 00:44:01.290 --> 00:44:03.090 to see what those doses do. 00:44:03.090 --> 00:44:04.950 And we had this in the last class. 00:44:04.950 --> 00:44:09.570 We talked about why that model is only defensibly linear, 00:44:09.570 --> 00:44:13.190 and why we don't have a fancier model for it. 00:44:13.190 --> 00:44:15.830 AUDIENCE: So we have the 230,000 deaths from a major accident, 00:44:15.830 --> 00:44:17.850 and I guess, how do we get that like a-- 00:44:17.850 --> 00:44:20.987 can we give that as a per year? 00:44:20.987 --> 00:44:22.070 R. SCOTT KEMP: Well, yeah. 00:44:22.070 --> 00:44:27.370 You could assume that the deaths expose people 00:44:27.370 --> 00:44:29.690 according to the decay, so that most people are 00:44:29.690 --> 00:44:31.330 exposed right away. 00:44:31.330 --> 00:44:34.610 But then there's some time between when those people are 00:44:34.610 --> 00:44:37.230 exposed and when they become-- when they die from cancer. 00:44:37.230 --> 00:44:39.290 And that could be 10 years. 00:44:39.290 --> 00:44:42.420 So there's this smearing-- 00:44:42.420 --> 00:44:43.685 a delay and smearing. 00:44:43.685 --> 00:44:45.060 AUDIENCE: [INAUDIBLE] get rid of. 00:44:45.060 --> 00:44:46.643 Like, if we were trying-- if we really 00:44:46.643 --> 00:44:48.100 wanted a per year [INAUDIBLE]. 00:44:48.100 --> 00:44:52.130 R. SCOTT KEMP: Order of magnitude from 1-- 00:44:52.130 --> 00:44:53.880 AUDIENCE: [INAUDIBLE] like over 100 years. 00:44:53.880 --> 00:44:55.380 But like, obviously, you can't [? divide by-- ?] 00:44:55.380 --> 00:44:56.020 R. SCOTT KEMP: So it depends. 00:44:56.020 --> 00:44:57.900 If you're assuming there's a constant rate of accidents, 00:44:57.900 --> 00:44:59.360 or are you assuming one accident? 00:44:59.360 --> 00:45:01.527 AUDIENCE: If we're trying to say there's a certain-- 00:45:01.527 --> 00:45:04.000 how people do [INAUDIBLE]-- totally other thing. 00:45:04.000 --> 00:45:08.580 But if you have an accident rate per year, what would 00:45:08.580 --> 00:45:10.855 that relative [INAUDIBLE]? 00:45:10.855 --> 00:45:12.340 We're trying to get that-- 00:45:12.340 --> 00:45:12.980 R. SCOTT KEMP: Then we're going to get 00:45:12.980 --> 00:45:14.640 to an accident rate per year. 00:45:14.640 --> 00:45:16.598 So maybe it will answer your question. 00:45:16.598 --> 00:45:17.140 AUDIENCE: OK. 00:45:17.140 --> 00:45:17.620 R. SCOTT KEMP: Yeah. 00:45:17.620 --> 00:45:18.140 Yeah? 00:45:18.140 --> 00:45:19.848 AUDIENCE: Maybe this is a naive question. 00:45:19.848 --> 00:45:23.380 But if you can't differentiate, how do you 00:45:23.380 --> 00:45:26.058 pick a background cancer rate? 00:45:26.058 --> 00:45:28.100 R. SCOTT KEMP: The background cancer rate is just 00:45:28.100 --> 00:45:31.540 the number of people who live-- 00:45:31.540 --> 00:45:33.560 who die versus the number of people-- 00:45:33.560 --> 00:45:33.580 AUDIENCE: [INAUDIBLE]. 00:45:33.580 --> 00:45:34.540 So it has nothing to do with how you 00:45:34.540 --> 00:45:35.990 delineate where it came from. 00:45:35.990 --> 00:45:36.740 R. SCOTT KEMP: No. 00:45:36.740 --> 00:45:37.240 Yeah. 00:45:37.240 --> 00:45:39.460 It's just the number of people who die from cancer. 00:45:39.460 --> 00:45:39.960 Yeah. 00:45:39.960 --> 00:45:42.690 Yeah? 00:45:42.690 --> 00:45:44.440 AUDIENCE: I guess it maybe kind of relates 00:45:44.440 --> 00:45:47.000 to previous questions, but how do 00:45:47.000 --> 00:45:50.180 we put this in perspective related to the background? 00:45:50.180 --> 00:45:53.240 Because if we look at 10 kilobecquerels per square meters 00:45:53.240 --> 00:45:55.120 times roughly 10 microsieverts per, 00:45:55.120 --> 00:45:58.332 it's like 1 millisievert per year that you get from this. 00:45:58.332 --> 00:46:00.040 I quickly looked up, what's the variation 00:46:00.040 --> 00:46:01.680 in background rate in Europe. 00:46:01.680 --> 00:46:04.895 It's anywhere between 1.5 to 5.8 millisieverts. 00:46:04.895 --> 00:46:07.020 It's like all of this disappears in the background. 00:46:07.020 --> 00:46:07.800 R. SCOTT KEMP: That's correct. 00:46:07.800 --> 00:46:09.217 AUDIENCE: [INAUDIBLE] doesn't show 00:46:09.217 --> 00:46:11.440 that the Netherlands, which apparently has the lowest 00:46:11.440 --> 00:46:14.580 background radiation, has significantly less cancer. 00:46:14.580 --> 00:46:16.640 So is the problem, then, that we're 00:46:16.640 --> 00:46:20.680 kind of including radiation effects in the relative risk 00:46:20.680 --> 00:46:22.940 background, and that's why everything is here? 00:46:22.940 --> 00:46:27.120 But why don't we see in LNT a more drastic change 00:46:27.120 --> 00:46:29.820 in cancer rates than between the Czech Republic, 00:46:29.820 --> 00:46:32.080 [INAUDIBLE] is 5.8 millisieverts, 00:46:32.080 --> 00:46:37.820 and then the Netherlands is 1.5, if 0.1 millisieverts per year 00:46:37.820 --> 00:46:41.180 creates such a drastic difference [INAUDIBLE]. 00:46:41.180 --> 00:46:45.540 R. SCOTT KEMP: So I think you're asking-- 00:46:45.540 --> 00:46:48.120 regional variation and background can be, let's say, 00:46:48.120 --> 00:46:50.860 factors of 2, factors of 3. 00:46:50.860 --> 00:46:54.780 Certain very small regions are higher. 00:46:54.780 --> 00:46:58.140 And yet, you don't see factors of 2 difference in the cancer 00:46:58.140 --> 00:47:00.160 rate in those regions. 00:47:00.160 --> 00:47:03.380 That's because most cancers are not caused by radiation. 00:47:03.380 --> 00:47:07.340 Most cancers are caused by just oxidation stress in the body, 00:47:07.340 --> 00:47:09.620 by crap you're eating, by whether you're 00:47:09.620 --> 00:47:11.020 smoking cigarettes. 00:47:11.020 --> 00:47:13.900 So already, the radiation-induced cancers 00:47:13.900 --> 00:47:16.460 are already in the noise. 00:47:16.460 --> 00:47:21.540 These are also in the noise, but they're not there. 00:47:21.540 --> 00:47:25.140 They are in the noise, but they're there. 00:47:25.140 --> 00:47:29.540 So it is incorrect to say, oh, well, lots of people 00:47:29.540 --> 00:47:30.240 die from cancer. 00:47:30.240 --> 00:47:32.220 Therefore, we should ignore this. 00:47:32.220 --> 00:47:38.990 Well, this accident did kill an additional 52,000 people. 00:47:38.990 --> 00:47:40.670 That happened. 00:47:40.670 --> 00:47:45.010 Now, you can-- or in this case, quarter million people 00:47:45.010 --> 00:47:48.360 once we do all the calculation. 00:47:48.360 --> 00:47:53.630 Yes, the background is noisy and big, and everyone dies. 00:47:53.630 --> 00:47:57.590 But to say that it's insignificant 00:47:57.590 --> 00:48:00.550 is incorrect, if I can take that same energy source 00:48:00.550 --> 00:48:04.350 and replace it with something that doesn't cause any deaths, 00:48:04.350 --> 00:48:07.130 or many fewer deaths. 00:48:07.130 --> 00:48:12.450 So it doesn't go away just because you can't see it. 00:48:15.190 --> 00:48:18.150 Because someone doesn't show up in the hospital, and they go, 00:48:18.150 --> 00:48:20.910 oh, this is a Chernobyl cancer. 00:48:20.910 --> 00:48:24.602 If we could do that, if we had the science-- 00:48:24.602 --> 00:48:26.810 everyone shows up at oncology, and they go, oh, yeah, 00:48:26.810 --> 00:48:31.870 this is from Fukushima, then the news reports would say, 00:48:31.870 --> 00:48:35.690 another 10,000 people died from Fukushima this year, 00:48:35.690 --> 00:48:39.610 and people would go crazy. 00:48:39.610 --> 00:48:41.490 It's the fact that we get to hide them 00:48:41.490 --> 00:48:44.790 that we don't have to take accountability for them, 00:48:44.790 --> 00:48:46.265 but they're still there. 00:48:46.265 --> 00:48:46.890 AUDIENCE: Yeah. 00:48:46.890 --> 00:48:48.307 I mean, the question was more, why 00:48:48.307 --> 00:48:51.190 does an LNT show a difference in background between countries 00:48:51.190 --> 00:48:52.190 if it's so much smaller? 00:48:52.190 --> 00:48:54.690 R. SCOTT KEMP: But the answer is because most of the cancers 00:48:54.690 --> 00:48:55.730 are not from radiation. 00:48:55.730 --> 00:48:57.240 Yeah. 00:48:57.240 --> 00:48:58.990 AUDIENCE: I was just going to [INAUDIBLE]. 00:48:58.990 --> 00:49:03.770 Do you use similar models to analyze the amount of deaths 00:49:03.770 --> 00:49:05.430 related to cancer or other factors-- 00:49:05.430 --> 00:49:10.530 respiratory or cardiovascular or fossil fuels 00:49:10.530 --> 00:49:13.490 and natural gas and other types of energy sources? 00:49:13.490 --> 00:49:16.170 Or they would you use a different type of model. 00:49:16.170 --> 00:49:19.090 R. SCOTT KEMP: You can only do such an easy model 00:49:19.090 --> 00:49:22.690 if there's a linear relationship between exposure 00:49:22.690 --> 00:49:25.330 and pathogenesis. 00:49:25.330 --> 00:49:28.970 So for a lot of diseases, which are not stochastic, 00:49:28.970 --> 00:49:31.050 which don't have a linear relationship, 00:49:31.050 --> 00:49:33.620 it will depend on thresholds, or there will 00:49:33.620 --> 00:49:35.160 be some nonlinear curvature. 00:49:35.160 --> 00:49:37.118 And that means you're going to have a much more 00:49:37.118 --> 00:49:38.765 sophisticated calculation of-- 00:49:38.765 --> 00:49:39.640 AUDIENCE: [INAUDIBLE] 00:49:39.640 --> 00:49:40.420 R. SCOTT KEMP: Yeah, exactly. 00:49:40.420 --> 00:49:42.940 How many people from this plant, and how many people live 00:49:42.940 --> 00:49:45.500 downwind of that particular plant of coal plant 00:49:45.500 --> 00:49:48.640 or something, and breathe that PM2.5. 00:49:48.640 --> 00:49:50.940 And yeah, so it's a lot more complicated. 00:49:50.940 --> 00:49:55.780 We actually get off easy because of the LNT model 00:49:55.780 --> 00:49:57.160 in doing this calculation. 00:49:57.160 --> 00:49:59.340 Yeah. 00:49:59.340 --> 00:49:59.840 All right. 00:49:59.840 --> 00:50:01.440 So let me keep going. 00:50:04.260 --> 00:50:08.280 So we've added the iodine, and we have this global dose, 00:50:08.280 --> 00:50:15.820 and the question is, do we believe Scott's number? 00:50:15.820 --> 00:50:16.900 I mean-- 00:50:16.900 --> 00:50:18.140 AUDIENCE: I will say no. 00:50:18.140 --> 00:50:19.120 No, I don't. 00:50:19.120 --> 00:50:23.080 But I take umbrage with the linear no-threshold model. 00:50:23.080 --> 00:50:26.020 I'm more in favor of the linear threshold model. 00:50:26.020 --> 00:50:28.620 R. SCOTT KEMP: Did you come to class? 00:50:28.620 --> 00:50:29.900 AUDIENCE: Yeah. 00:50:29.900 --> 00:50:31.240 Well, I wasn't there last time. 00:50:31.240 --> 00:50:32.360 I was [INAUDIBLE]. 00:50:32.360 --> 00:50:33.440 R. SCOTT KEMP: OK. 00:50:33.440 --> 00:50:35.200 AUDIENCE: But the-- 00:50:35.200 --> 00:50:36.747 I mean, [INAUDIBLE]. 00:50:36.747 --> 00:50:37.580 R. SCOTT KEMP: Yeah. 00:50:37.580 --> 00:50:39.820 So this is where things start to go wrong. 00:50:39.820 --> 00:50:41.500 People are like, oh, I hate this number. 00:50:41.500 --> 00:50:43.840 Now I'm going to choose not to believe 00:50:43.840 --> 00:50:47.240 the linear no-threshold model because it saves me from having 00:50:47.240 --> 00:50:48.320 to deal with this number. 00:50:48.320 --> 00:50:48.820 I'm sorry. 00:50:48.820 --> 00:50:50.680 That's not scientific. 00:50:50.680 --> 00:50:53.280 You don't just get to pick a model that is not statistically 00:50:53.280 --> 00:50:56.400 defensible because you don't like the result. 00:50:56.400 --> 00:51:00.160 So I appreciate that people do this all the time, 00:51:00.160 --> 00:51:04.140 but we're not going to do that here. 00:51:04.140 --> 00:51:04.640 All right. 00:51:04.640 --> 00:51:06.480 Two more questions, then I need to move on. 00:51:06.480 --> 00:51:08.480 AUDIENCE: I was going to say, does it really matter? 00:51:08.480 --> 00:51:10.688 If we know-- if there was some theoretical threshold, 00:51:10.688 --> 00:51:14.382 if it was really small, how much does that really affect the-- 00:51:14.382 --> 00:51:16.840 would it be magnitudes, or if it's a threshold that small-- 00:51:16.840 --> 00:51:21.453 R. SCOTT KEMP: It could have a significant effect. 00:51:21.453 --> 00:51:23.120 I don't know if magnitudes, but it could 00:51:23.120 --> 00:51:24.500 have a significant effect. 00:51:24.500 --> 00:51:26.560 But the argument, the mechanistic argument 00:51:26.560 --> 00:51:29.380 against the threshold, remember, is the fact 00:51:29.380 --> 00:51:31.740 that in low LET radiation transport, 00:51:31.740 --> 00:51:34.360 you have Compton scattering and photoelectric effect. 00:51:34.360 --> 00:51:35.540 So you have these chains. 00:51:35.540 --> 00:51:37.040 You can create double strand breaks. 00:51:37.040 --> 00:51:39.980 Double strand breaks are basically difficult to repair. 00:51:39.980 --> 00:51:43.940 A single photon can cause a poorly repaired double strand 00:51:43.940 --> 00:51:46.180 break that then causes cancer. 00:51:46.180 --> 00:51:49.820 As long as that chain of physics is possible, 00:51:49.820 --> 00:51:54.300 there cannot be a threshold. 00:51:54.300 --> 00:51:58.880 So I don't see any basis for believing there's a threshold. 00:51:58.880 --> 00:52:01.300 It's neither empirically defensible 00:52:01.300 --> 00:52:03.040 nor mechanistically defensible. 00:52:05.706 --> 00:52:06.640 Yes? 00:52:06.640 --> 00:52:09.360 AUDIENCE: With such a marginal change from background 00:52:09.360 --> 00:52:11.460 having [INAUDIBLE] effect, I wonder 00:52:11.460 --> 00:52:17.620 how this would stack up against other industries that 00:52:17.620 --> 00:52:19.260 cause harm, arguably. 00:52:19.260 --> 00:52:22.100 How would this compare to the increase in deaths caused 00:52:22.100 --> 00:52:23.870 by McDonald's, [INAUDIBLE]? 00:52:23.870 --> 00:52:24.620 R. SCOTT KEMP: Oh. 00:52:24.620 --> 00:52:25.975 Probably-- yeah. 00:52:25.975 --> 00:52:26.600 AUDIENCE: Tiny. 00:52:26.600 --> 00:52:27.683 R. SCOTT KEMP: Tiny, yeah. 00:52:27.683 --> 00:52:30.070 No, there are-- driving your car, 00:52:30.070 --> 00:52:33.510 eating peanut butter, which contains aflatoxin, 00:52:33.510 --> 00:52:36.350 choosing to walk on the side of the road versus not walking 00:52:36.350 --> 00:52:39.470 on the side of the road, wearing a bicycle helmet-- these things 00:52:39.470 --> 00:52:42.750 have huge impacts on mortality. 00:52:42.750 --> 00:52:43.830 Yeah? 00:52:43.830 --> 00:52:48.710 AUDIENCE: So your point isn't that this number should then 00:52:48.710 --> 00:52:51.910 be used to justify or not justify nuclear, but instead, 00:52:51.910 --> 00:52:53.070 be used for comparison. 00:52:53.070 --> 00:52:54.670 R. SCOTT KEMP: Only that. 00:52:54.670 --> 00:52:56.870 You can only take this and compare it 00:52:56.870 --> 00:53:00.630 to other technologies which are also killing you in other ways. 00:53:00.630 --> 00:53:02.510 That's all you can use this for. 00:53:02.510 --> 00:53:03.110 Yeah. 00:53:03.110 --> 00:53:06.310 This is not-- because 250,000 feels big, 00:53:06.310 --> 00:53:07.930 it's not an argument against nuclear. 00:53:07.930 --> 00:53:09.810 Let's be really clear about that. 00:53:09.810 --> 00:53:13.190 It is not an argument against nuclear. 00:53:13.190 --> 00:53:17.070 And in the last half hour, I want to get to that comparison. 00:53:17.070 --> 00:53:18.497 Yeah? 00:53:18.497 --> 00:53:20.830 AUDIENCE: This is really only for comparing other energy 00:53:20.830 --> 00:53:21.330 forms. 00:53:21.330 --> 00:53:22.750 We can't compare this number to-- 00:53:22.750 --> 00:53:24.790 R. SCOTT KEMP: Peanut butter. 00:53:24.790 --> 00:53:25.290 Correct. 00:53:25.290 --> 00:53:26.950 Because they're not a substitution. 00:53:26.950 --> 00:53:30.650 But we can compare it to other forms of electricity generation, 00:53:30.650 --> 00:53:32.530 and that's what we should be doing. 00:53:32.530 --> 00:53:34.090 So let's do it. 00:53:34.090 --> 00:53:38.450 So first, I want to talk a little bit about whether we 00:53:38.450 --> 00:53:39.310 believe this number. 00:53:39.310 --> 00:53:41.690 Because how many people have heard 4,000 people died from 00:53:41.690 --> 00:53:43.790 Chernobyl? 00:53:43.790 --> 00:53:44.290 Yeah. 00:53:44.290 --> 00:53:44.870 OK. 00:53:44.870 --> 00:53:47.050 Where did that come from? 00:53:47.050 --> 00:53:50.330 So here are the published estimates-- 00:53:50.330 --> 00:53:53.390 4,000 deaths, 9,000 deaths, 16,000 deaths, 40,000 deaths, 00:53:53.390 --> 00:53:56.890 60,000 deaths, 270,000 deaths. 00:53:56.890 --> 00:54:00.410 But you have to look at what is being assumed. 00:54:00.410 --> 00:54:03.190 Only the most exposed cohort of people, 00:54:03.190 --> 00:54:04.890 and only in three countries. 00:54:04.890 --> 00:54:07.350 Same people-- this is where the 4,000 deaths come from. 00:54:07.350 --> 00:54:11.110 It was a study done by the UN Chernobyl people. 00:54:11.110 --> 00:54:12.570 It was actually done pretty shortly 00:54:12.570 --> 00:54:15.450 after it was reported in 2005. 00:54:15.450 --> 00:54:18.510 Then they said, oh, we'll look at all people, 00:54:18.510 --> 00:54:20.990 not just like the most exposed populations. 00:54:20.990 --> 00:54:23.620 And then that number went up to 9,000. 00:54:23.620 --> 00:54:27.220 Then the World Health Organization research on cancer 00:54:27.220 --> 00:54:31.020 came up with a number up until 2065. 00:54:31.020 --> 00:54:35.180 Only Europe of 16,000. 00:54:35.180 --> 00:54:38.340 A group of independent scientists out of the UK 00:54:38.340 --> 00:54:40.700 didn't like the stuff that the UN 00:54:40.700 --> 00:54:44.340 was doing, with only reporting subsets of the population, 00:54:44.340 --> 00:54:46.400 and only the people in these three countries. 00:54:46.400 --> 00:54:48.360 And they felt this was unfair. 00:54:48.360 --> 00:54:50.260 So they produced these reports. 00:54:50.260 --> 00:54:52.440 One in 2006, and one in 2016-- 00:54:52.440 --> 00:54:54.540 40,000 and 60,000. 00:54:54.540 --> 00:54:56.480 This one is still limited to Europe. 00:54:56.480 --> 00:54:58.140 This one is global. 00:54:58.140 --> 00:55:01.860 And then you have this Greenpeace report-- 00:55:01.860 --> 00:55:03.900 270,000. 00:55:03.900 --> 00:55:08.460 So all of these credible people come out less. 00:55:08.460 --> 00:55:14.140 Roughly, one fourth to one fifth of the number 00:55:14.140 --> 00:55:16.100 that I came up with. 00:55:16.100 --> 00:55:18.820 So why is that? 00:55:18.820 --> 00:55:23.960 So the answer is that they use a 1980s US Department of Energy 00:55:23.960 --> 00:55:27.120 report on the contamination to dose factor. 00:55:27.120 --> 00:55:28.300 That is, this factor. 00:55:28.300 --> 00:55:29.260 Remember this? 00:55:29.260 --> 00:55:31.440 And I said this was controversial. 00:55:31.440 --> 00:55:36.320 They used a number that reads 0.2 here. 00:55:36.320 --> 00:55:39.840 That 1980s report doesn't have any justification for that 00:55:39.840 --> 00:55:40.340 number. 00:55:40.340 --> 00:55:42.560 They just tabulate numbers for a bunch of isotopes, 00:55:42.560 --> 00:55:44.400 and that's it. 00:55:44.400 --> 00:55:46.800 So why did I use this one? 00:55:46.800 --> 00:55:52.640 I used this one because after Fukushima, the French Institute 00:55:52.640 --> 00:55:58.920 for Radiation Protection went to the Fukushima site using 00:55:58.920 --> 00:56:03.160 an energy-resolving radiation detector, 00:56:03.160 --> 00:56:05.520 like a high purity germanium detector. 00:56:05.520 --> 00:56:09.680 And they measured the photon flux off the ground. 00:56:09.680 --> 00:56:14.080 And then they measured the contamination on the ground, 00:56:14.080 --> 00:56:15.260 and they calculated it. 00:56:15.260 --> 00:56:20.300 And the number that they got was 0.86. 00:56:20.300 --> 00:56:23.220 It basically matches this number. 00:56:23.220 --> 00:56:26.160 And that's the most recent scientific validation of this. 00:56:26.160 --> 00:56:30.560 So I use this number because it's been validated twice, 00:56:30.560 --> 00:56:32.340 and we know how it was done. 00:56:32.340 --> 00:56:36.980 And the other number that people have been using for this report 00:56:36.980 --> 00:56:39.740 seems to come out of nowhere. 00:56:39.740 --> 00:56:43.660 Now, here's the caveat. 00:56:43.660 --> 00:56:47.460 That measurement was done four months after the Fukushima 00:56:47.460 --> 00:56:48.980 accident. 00:56:48.980 --> 00:56:52.500 So it is possible, in fact, likely, 00:56:52.500 --> 00:56:55.940 that there is a kind of washing away effect, 00:56:55.940 --> 00:57:00.300 where stuff slowly settles into the ground 00:57:00.300 --> 00:57:02.700 and becomes less available. 00:57:02.700 --> 00:57:05.500 And therefore, the dose slowly decays 00:57:05.500 --> 00:57:10.260 from physiochemical changes in the soil. 00:57:10.260 --> 00:57:14.900 And so we would expect, actually, some kind of decay 00:57:14.900 --> 00:57:17.060 for this. 00:57:17.060 --> 00:57:20.390 So unfortunately, there isn't any good studies, 00:57:20.390 --> 00:57:23.870 and it's all extremely dependent on the soil 00:57:23.870 --> 00:57:26.950 chemistry of where you live. 00:57:26.950 --> 00:57:31.430 But there is one datum taken six years after the Chernobyl 00:57:31.430 --> 00:57:34.270 accident where they did this dose estimation. 00:57:34.270 --> 00:57:38.190 And if I take that one datum, and I fit an exponential decay 00:57:38.190 --> 00:57:41.910 model to it, then what I would estimate 00:57:41.910 --> 00:57:45.630 is that the cesium contribution would actually 00:57:45.630 --> 00:57:49.270 be about a fifth of what we calculated, 00:57:49.270 --> 00:57:54.350 which explains exactly this differential integrated 00:57:54.350 --> 00:57:55.910 over the lifetime. 00:57:55.910 --> 00:57:59.790 And therefore, if we go back, and we 00:57:59.790 --> 00:58:02.830 take this number and this number, 00:58:02.830 --> 00:58:08.890 and we multiply that times a fifth, you get a smaller value. 00:58:08.890 --> 00:58:10.810 Let's see if I have it in my notes here. 00:58:27.410 --> 00:58:33.490 Oh, it's about 60% total dose change, since a lot of this dose 00:58:33.490 --> 00:58:36.530 comes from iodine which doesn't have that effect. 00:58:36.530 --> 00:58:41.610 So somewhere between 60% of a 230,000. 00:58:41.610 --> 00:58:46.370 And 230,000 is the number that I think we should go with. 00:58:46.370 --> 00:58:50.230 But that is why there's a lot of discrepancy in the literature, 00:58:50.230 --> 00:58:56.600 is about how you assume the dose response would be, 00:58:56.600 --> 00:59:00.510 or the contamination to dose would be. 00:59:00.510 --> 00:59:03.870 So now let's put some-- let's convert this into money. 00:59:03.870 --> 00:59:07.650 So we'll have this in mind that-- 00:59:07.650 --> 00:59:09.730 I'll just write it here-- 00:59:09.730 --> 00:59:22.375 somewhere between 230,000 and 0.6 times 230,000 deaths. 00:59:22.375 --> 00:59:23.250 I'll use this number. 00:59:23.250 --> 00:59:25.042 And then at the end, I'll put this back in. 00:59:27.790 --> 00:59:33.190 So how often do these accidents occur? 00:59:33.190 --> 00:59:38.630 So one of the problems is that we don't really know. 00:59:38.630 --> 00:59:41.870 And one way that this is often done is using PRA. 00:59:41.870 --> 00:59:44.785 And I already talked a little bit about PRA in this class. 00:59:44.785 --> 00:59:47.410 It's a bottom-up approach where you build these decision trees. 00:59:47.410 --> 00:59:48.910 And then for every kind of event you 00:59:48.910 --> 00:59:52.670 have to have some probability that you assign to it. 00:59:52.670 --> 00:59:54.070 And one of the problems with this 00:59:54.070 --> 00:59:56.810 is that, well, you need to know a lot of things. 00:59:56.810 --> 00:59:59.750 You need to know every branching probability. 00:59:59.750 --> 01:00:02.030 And you probably have never observed most 01:00:02.030 --> 01:00:04.150 of those branches happening. 01:00:04.150 --> 01:00:07.050 And so you're kind of like guessing at it. 01:00:07.050 --> 01:00:09.910 And it's a little bit of a garbage in, garbage out model. 01:00:09.910 --> 01:00:11.830 Plus, you have to decide which trees 01:00:11.830 --> 01:00:13.560 you're going to put on here. 01:00:13.560 --> 01:00:15.120 And one of the problems is that a lot 01:00:15.120 --> 01:00:20.240 of historical big accidents have followed 01:00:20.240 --> 01:00:23.255 paths that were not in the PRA model at the time 01:00:23.255 --> 01:00:24.380 that the accident occurred. 01:00:24.380 --> 01:00:30.560 So we don't really have a priori belief that the PRA is complete. 01:00:30.560 --> 01:00:34.170 They call this-- so one of these is-- they call that model 01:00:34.170 --> 01:00:35.920 uncertainty, and they call the uncertainty 01:00:35.920 --> 01:00:39.020 with all these branching points, epistemic uncertainty. 01:00:39.020 --> 01:00:42.320 And there's a lot of uncertainty in these models. 01:00:42.320 --> 01:00:45.520 Nevertheless, the practitioners of PRA-- 01:00:45.520 --> 01:00:47.760 well, the careful practitioners of PRA 01:00:47.760 --> 01:00:50.760 would never say that you should take a PRA model 01:00:50.760 --> 01:00:52.400 and take the final result and use it 01:00:52.400 --> 01:00:57.020 as a measure of the safety of a reactor. 01:00:57.020 --> 01:00:58.960 All they will say is, well, if I change 01:00:58.960 --> 01:01:00.880 this thing versus changing that thing, what is 01:01:00.880 --> 01:01:03.000 my relative change in safety? 01:01:03.000 --> 01:01:05.960 And that's a totally valid and good use of PRA. 01:01:05.960 --> 01:01:09.960 But there's unfortunately this temptation 01:01:09.960 --> 01:01:14.520 to take the final number and treat it as a true number. 01:01:14.520 --> 01:01:21.780 And the NRC has recently adopted this view when they do back fit. 01:01:21.780 --> 01:01:24.580 And it makes sense for back fit, because maybe you're 01:01:24.580 --> 01:01:26.760 looking at things which are explicitly in the model, 01:01:26.760 --> 01:01:28.060 and you're looking about, how am I 01:01:28.060 --> 01:01:29.643 going to change some piece of hardware 01:01:29.643 --> 01:01:30.960 to change this value here? 01:01:30.960 --> 01:01:34.380 So you can justify the use of PRA back fit 01:01:34.380 --> 01:01:37.620 from the perspective of relative safety change. 01:01:37.620 --> 01:01:39.540 But you can't if you're going to do 01:01:39.540 --> 01:01:42.660 a cost-benefit analysis, which then 01:01:42.660 --> 01:01:46.100 assumes a certain absolute accident risk. 01:01:46.100 --> 01:01:49.460 That's the problem. 01:01:49.460 --> 01:01:52.640 And we have now this discussion of risk-informed licensing, 01:01:52.640 --> 01:01:56.220 where this is what people want to do is take this approach 01:01:56.220 --> 01:01:58.580 and assume it to be the true measure of safety 01:01:58.580 --> 01:01:59.720 of an accident. 01:01:59.720 --> 01:02:05.460 And that's not a good choice because essentially, we 01:02:05.460 --> 01:02:07.580 are favoring a model that we know 01:02:07.580 --> 01:02:11.030 to be a, a model, b, incomplete, and c, 01:02:11.030 --> 01:02:14.630 full of inputs which we are highly uncertain of, 01:02:14.630 --> 01:02:19.070 in favor of empirical data. 01:02:19.070 --> 01:02:21.350 In Occam's razor, which we talk about, 01:02:21.350 --> 01:02:24.310 would say any time you have a model with enormous numbers 01:02:24.310 --> 01:02:27.790 of inputs, you should be biased against it 01:02:27.790 --> 01:02:34.670 because it's probably missing something or leading you astray. 01:02:34.670 --> 01:02:37.870 Let's look at the results produced by these models. 01:02:37.870 --> 01:02:39.810 So here's the core damage frequency, 01:02:39.810 --> 01:02:42.090 and this is the large, early release frequency. 01:02:42.090 --> 01:02:45.350 This is the closest thing we have to a big nuclear reactor 01:02:45.350 --> 01:02:46.270 accident. 01:02:46.270 --> 01:02:50.790 And these are for results for AP 1,000, the EPR in Europe. 01:02:50.790 --> 01:02:53.490 There's the Chinese-- I'm going to put a Gen 4 reactor in there. 01:02:53.490 --> 01:02:55.448 This is HTPM from China. 01:02:55.448 --> 01:02:56.990 And then here's the models for these. 01:02:56.990 --> 01:02:58.810 These are not really widely published. 01:02:58.810 --> 01:03:01.610 You have to talk to people to get these numbers-- operating 01:03:01.610 --> 01:03:04.746 VWRs and PWRs. 01:03:04.746 --> 01:03:07.890 And so you can see what the results are. 01:03:07.890 --> 01:03:10.390 EPR, I think they're really doing a good job. 01:03:10.390 --> 01:03:13.250 They're basically just saying, we 01:03:13.250 --> 01:03:16.590 don't really know-- something less than 10 to the minus 6. 01:03:16.590 --> 01:03:19.970 Incidentally, EPR has a lot of fancy safety systems compared 01:03:19.970 --> 01:03:23.590 to AP-1000, but it's much more expensive. 01:03:23.590 --> 01:03:25.450 But AP-1000 people-- Westinghouse 01:03:25.450 --> 01:03:27.450 is willing to put two digits of significance 01:03:27.450 --> 01:03:31.730 and claim this enormous exponent of safety. 01:03:31.730 --> 01:03:35.350 And I think that this is a pretty dubious claim. 01:03:35.350 --> 01:03:37.470 I think EPR people are doing better at it. 01:03:40.330 --> 01:03:44.790 So if you believe these numbers are the true measure of safety-- 01:03:44.790 --> 01:03:45.410 yep? 01:03:45.410 --> 01:03:46.827 AUDIENCE: I may have miscaught it. 01:03:46.827 --> 01:03:49.170 What's the unit frequency, from what to what? 01:03:49.170 --> 01:03:51.490 R. SCOTT KEMP: These are events per year. 01:03:51.490 --> 01:03:52.350 AUDIENCE: Per year. 01:03:52.350 --> 01:03:52.850 Thank you. 01:03:52.850 --> 01:03:53.683 R. SCOTT KEMP: Yeah. 01:03:53.683 --> 01:03:56.070 So 10 to the minus events-- eight events per year. 01:03:59.050 --> 01:04:01.130 So if you assume this to be correct, 01:04:01.130 --> 01:04:03.330 we would say how many large releases 01:04:03.330 --> 01:04:05.350 would we get over 50 years? 01:04:05.350 --> 01:04:11.550 Assuming 80% of global electricity came from nuclear, 01:04:11.550 --> 01:04:15.470 and we have a linear growth from where we are now 01:04:15.470 --> 01:04:18.910 to get to that in 50 years. 01:04:18.910 --> 01:04:20.850 And so I've done the integral for you. 01:04:20.850 --> 01:04:23.590 And look, I mean, basically not zero. 01:04:23.590 --> 01:04:25.150 You would have no accidents, even 01:04:25.150 --> 01:04:28.830 though 80% of the power in the world came from nuclear. 01:04:28.830 --> 01:04:31.430 Here is less than 0.4. 01:04:31.430 --> 01:04:34.630 Only for existing operating BWR would 01:04:34.630 --> 01:04:39.970 you have more than one accident per year. 01:04:44.070 --> 01:04:47.590 If you're in this regime, you would have a problem. 01:04:47.590 --> 01:04:50.190 Can you imagine having a large nuclear reactor accident 01:04:50.190 --> 01:04:51.410 every single year? 01:04:51.410 --> 01:04:54.930 AUDIENCE: [INAUDIBLE] 50 years? 01:04:54.930 --> 01:04:57.100 R. SCOTT KEMP: Oh. 01:04:57.100 --> 01:04:57.600 Yes. 01:05:02.310 --> 01:05:03.410 Yes, you're right. 01:05:03.410 --> 01:05:03.550 Yeah. 01:05:03.550 --> 01:05:04.883 So this would be-- that's right. 01:05:04.883 --> 01:05:06.900 This would be roughly one every 50 years. 01:05:06.900 --> 01:05:09.520 So that's basically where we are now. 01:05:09.520 --> 01:05:11.660 So that would be fine. 01:05:11.660 --> 01:05:13.048 That would be fine. 01:05:13.048 --> 01:05:14.840 Well, I don't know about fine, but that's-- 01:05:14.840 --> 01:05:15.960 AUDIENCE: It's not bad. 01:05:15.960 --> 01:05:16.380 R. SCOTT KEMP: It's not bad. 01:05:16.380 --> 01:05:18.297 It's better than what we've seen historically. 01:05:18.297 --> 01:05:19.600 AUDIENCE: [INAUDIBLE] 01:05:19.600 --> 01:05:21.240 R. SCOTT KEMP: Yeah. 01:05:21.240 --> 01:05:24.040 So this if these models are correct, 01:05:24.040 --> 01:05:27.240 we really have nothing to fear from nuclear. 01:05:27.240 --> 01:05:29.720 But the problem is that the history is not 01:05:29.720 --> 01:05:30.700 comport with that. 01:05:30.700 --> 01:05:34.920 So these are the actual accidents, different INES 01:05:34.920 --> 01:05:35.420 scales. 01:05:35.420 --> 01:05:36.170 There's Chernobyl. 01:05:36.170 --> 01:05:37.800 There's Fukushima. 01:05:37.800 --> 01:05:40.340 The ones that are not black, the open circles, 01:05:40.340 --> 01:05:43.730 these are nonreactor fuel cycle accidents. 01:05:43.730 --> 01:05:45.480 So they are associated with nuclear power, 01:05:45.480 --> 01:05:47.400 but they are not reactors. 01:05:47.400 --> 01:05:52.360 So this is a Soviet fuel reprocessing site. 01:05:52.360 --> 01:05:54.760 These are also such sites. 01:05:54.760 --> 01:06:01.320 And you see that they also can contribute significantly. 01:06:01.320 --> 01:06:03.180 Another thing to see is that this basically 01:06:03.180 --> 01:06:04.280 kind of like a power law. 01:06:04.280 --> 01:06:09.880 We expect more accidents as we get to lower intensity; 01:06:09.880 --> 01:06:12.120 fewer accidents as we get to worse intensity. 01:06:12.120 --> 01:06:14.140 But it's not exactly a power law. 01:06:14.140 --> 01:06:16.700 There's only one INES 6. 01:06:16.700 --> 01:06:19.080 That's probably because of quantization noise. 01:06:19.080 --> 01:06:20.522 We're a little bit undersampling. 01:06:20.522 --> 01:06:21.980 And the other thing to point out is 01:06:21.980 --> 01:06:26.500 that the accident rate is decreasing, 01:06:26.500 --> 01:06:30.020 at least for category 4. 01:06:30.020 --> 01:06:31.560 So maybe things are getting safer. 01:06:31.560 --> 01:06:33.940 I would argue that they have been getting safer. 01:06:33.940 --> 01:06:36.860 We're better at safety. 01:06:36.860 --> 01:06:42.660 So what should we assume for these big accidents? 01:06:42.660 --> 01:06:45.100 So here's the thing. 01:06:45.100 --> 01:06:48.780 These large INES 7 accidents are typically 01:06:48.780 --> 01:06:52.780 what we call design beyond design basis accidents. 01:06:52.780 --> 01:06:57.260 That means we did not do things to the reactor to prevent 01:06:57.260 --> 01:06:59.180 these accidents from occurring. 01:06:59.180 --> 01:07:02.030 So although we are improving the reactor safety 01:07:02.030 --> 01:07:06.950 for small events which we are designing for, 01:07:06.950 --> 01:07:09.750 we are not doing anything to deal with these 01:07:09.750 --> 01:07:11.985 beyond design basis accidents. 01:07:11.985 --> 01:07:13.110 AUDIENCE: Well, that goes-- 01:07:13.110 --> 01:07:14.910 doesn't go for Chernobyl, though. 01:07:14.910 --> 01:07:19.750 You can't just pull out all your [INAUDIBLE]. 01:07:19.750 --> 01:07:22.427 I get it from Fukushima, but Chernobyl, I 01:07:22.427 --> 01:07:24.010 feel like it's kind of the equivalent, 01:07:24.010 --> 01:07:27.430 from an operator standpoint, of getting blackout wasted, 01:07:27.430 --> 01:07:30.057 driving your car into an [INAUDIBLE]. 01:07:30.057 --> 01:07:30.890 R. SCOTT KEMP: Yeah. 01:07:30.890 --> 01:07:32.050 But that's-- exactly. 01:07:32.050 --> 01:07:34.950 That's why it's beyond design basis accident. 01:07:34.950 --> 01:07:37.650 It's like the operator did something stupid. 01:07:37.650 --> 01:07:40.870 And there's nothing to say that some operator of the future 01:07:40.870 --> 01:07:43.230 won't also do something stupid. 01:07:46.340 --> 01:07:49.310 There's going to be all kinds of these things that haven't 01:07:49.310 --> 01:07:51.070 happened that you can imagine-- 01:07:51.070 --> 01:07:55.910 terrorist attacks, whatever-- that could 01:07:55.910 --> 01:07:59.690 result in intentional sabotage. 01:07:59.690 --> 01:08:02.770 So there are going to be these events, 01:08:02.770 --> 01:08:07.330 and it's going to be really hard to know what their frequency is. 01:08:07.330 --> 01:08:12.850 What I suggest is that in the absence of any good guidance 01:08:12.850 --> 01:08:16.770 about how often these things will occur, 01:08:16.770 --> 01:08:19.210 the best, most defensible thing we can do 01:08:19.210 --> 01:08:23.010 is just take the historical rate and say, 01:08:23.010 --> 01:08:30.050 that's what we've seen for these beyond design basis accidents. 01:08:30.050 --> 01:08:33.390 Anything else, you're injecting some additional information. 01:08:33.390 --> 01:08:38.170 And I don't know that we have any additional information. 01:08:38.170 --> 01:08:41.450 So here's the history of reactor. 01:08:41.450 --> 01:08:45.250 We have 21,500 reactor years of experience, 01:08:45.250 --> 01:08:50.930 and expectation over 50 years with no growth would be about 01:08:50.930 --> 01:08:55.472 two reactor accidents, which is kind of what we've seen. 01:08:55.472 --> 01:08:56.930 But if we did-- again, here's where 01:08:56.930 --> 01:08:58.689 I was going to make that point earlier. 01:08:58.689 --> 01:09:01.710 If we did have that growth to 80% electricity, 01:09:01.710 --> 01:09:06.990 we would be at 40 INES 7 events every 50 years, which 01:09:06.990 --> 01:09:09.590 would be roughly one per year, and then we 01:09:09.590 --> 01:09:12.350 would have a problem. 01:09:12.350 --> 01:09:16.290 So this is the problem with these events-- 01:09:16.290 --> 01:09:18.710 a major global expansion of nuclear power. 01:09:18.710 --> 01:09:23.350 Unless this beyond design basis number somehow comes down, 01:09:23.350 --> 01:09:25.910 it still presents an ongoing problem 01:09:25.910 --> 01:09:28.390 that we have to deal with that would basically 01:09:28.390 --> 01:09:30.410 kill the public acceptance of nuclear, 01:09:30.410 --> 01:09:35.130 I would argue, at roughly an accident a year-- 01:09:35.130 --> 01:09:37.670 giant accident a year. 01:09:37.670 --> 01:09:38.283 Yeah? 01:09:38.283 --> 01:09:39.950 AUDIENCE: Could we also argue that if we 01:09:39.950 --> 01:09:41.950 saw an increase in these types of accidents that 01:09:41.950 --> 01:09:44.590 were, like you were saying, beyond design 01:09:44.590 --> 01:09:48.229 basis, human error, and we start-- if we have those more 01:09:48.229 --> 01:09:51.710 often, would we not then be, I guess, 01:09:51.710 --> 01:09:55.948 incentivized to make that more safe more quickly. 01:09:55.948 --> 01:09:57.240 R. SCOTT KEMP: I would hope so. 01:09:57.240 --> 01:09:59.400 So we would have thought that after Fukushima, we 01:09:59.400 --> 01:10:00.940 would have made-- 01:10:00.940 --> 01:10:04.520 we would have added filter vents, and most of the world 01:10:04.520 --> 01:10:05.020 did. 01:10:05.020 --> 01:10:07.560 But the United States said no. 01:10:07.560 --> 01:10:09.307 So it depends. 01:10:09.307 --> 01:10:11.640 If it's happening every year, I think political pressure 01:10:11.640 --> 01:10:14.363 to do it would be a lot more. 01:10:14.363 --> 01:10:15.280 AUDIENCE: [INAUDIBLE]. 01:10:15.280 --> 01:10:17.960 I guess-- I hope we did learn not to site them in places that 01:10:17.960 --> 01:10:20.125 historically have tsunamis. 01:10:20.125 --> 01:10:21.500 R. SCOTT KEMP: Hopefully, we did. 01:10:21.500 --> 01:10:25.960 We'll find out. 01:10:25.960 --> 01:10:27.682 AUDIENCE: Probably not. 01:10:27.682 --> 01:10:29.140 R. SCOTT KEMP: This is the problem. 01:10:29.140 --> 01:10:30.820 I think memory is short. 01:10:30.820 --> 01:10:33.840 It's like there's some half life associated with learning 01:10:33.840 --> 01:10:35.680 from these events. 01:10:35.680 --> 01:10:38.180 In the immediate aftermath of Fukushima, 01:10:38.180 --> 01:10:39.780 everyone is angry at nuclear. 01:10:39.780 --> 01:10:40.580 Everyone hates it. 01:10:40.580 --> 01:10:42.420 Enrollments in nuclear engineering drop. 01:10:42.420 --> 01:10:46.800 And then three years later, everything creeps back up 01:10:46.800 --> 01:10:49.080 and people forget. 01:10:49.080 --> 01:10:50.980 So we'll see. 01:10:50.980 --> 01:10:52.880 So we can just take these numbers, 01:10:52.880 --> 01:10:59.540 and we can calculate a large early release frequency. 01:10:59.540 --> 01:11:01.980 I guess you can't really see that this is red. 01:11:01.980 --> 01:11:07.420 So if we take that rate, and we take the existing fleet, 01:11:07.420 --> 01:11:10.820 and we take the weighted average, 01:11:10.820 --> 01:11:14.080 we would see that we would be at 4-- sorry. 01:11:14.080 --> 01:11:17.380 This is the PRA result. We are at 2 times to the minus 4. 01:11:17.380 --> 01:11:19.780 The PRA result for the existing fleet 01:11:19.780 --> 01:11:22.740 says we should be at 4.5 times 10 to the minus 6. 01:11:22.740 --> 01:11:26.780 This is the evidence that the PRA result is incomplete. 01:11:26.780 --> 01:11:29.620 PRA has said the reactors operating 01:11:29.620 --> 01:11:33.380 today had this accident rate, and this is the rate 01:11:33.380 --> 01:11:35.840 that we've actually observed. 01:11:35.840 --> 01:11:41.180 It's a factor of 50 times difference. 01:11:41.180 --> 01:11:45.580 So the PRA is only capturing a small fraction 01:11:45.580 --> 01:11:47.920 of the total risk. 01:11:47.920 --> 01:11:50.740 Let me see if I can blast through and finish this up. 01:11:50.740 --> 01:11:55.900 So here, we can pick-- various is the NRC safety target. 01:11:55.900 --> 01:11:59.200 Here's a historical accident rate for all events over 5 01:11:59.200 --> 01:12:01.340 to match the safety target. 01:12:01.340 --> 01:12:05.000 Here's the true historical rate for only INES 7. 01:12:05.000 --> 01:12:06.860 If you go back and look at this plot-- 01:12:09.840 --> 01:12:13.880 if you look at these other events, pretty tiny releases-- 01:12:13.880 --> 01:12:18.880 0.2 petabecquerel, 0.013 petabecquerel cesium-- 01:12:18.880 --> 01:12:19.960 tiny. 01:12:19.960 --> 01:12:23.380 Only these big things beyond design basis matter. 01:12:23.380 --> 01:12:25.440 So my suggestion is we just ignore 01:12:25.440 --> 01:12:26.620 all these other accidents. 01:12:26.620 --> 01:12:27.720 They don't matter. 01:12:27.720 --> 01:12:31.240 Only things that matter are the beyond design basis accidents. 01:12:31.240 --> 01:12:37.600 And so we'll take only the beyond design the INES 7 rate 01:12:37.600 --> 01:12:40.660 as our accident rate, and we'll ignore all these others. 01:12:40.660 --> 01:12:44.080 And that is 10 to the minus 4 per reactor year. 01:12:44.080 --> 01:12:47.280 And if we assume that every reactor is roughly a gigawatt, 01:12:47.280 --> 01:12:52.370 it's 10 to the minus 11 per megawatt hour. 01:12:52.370 --> 01:12:55.070 And then we can take our number of deaths-- 01:12:57.670 --> 01:13:03.610 so 10-- we want terawatt hours, which 01:13:03.610 --> 01:13:07.010 is 10 to the 12 watt hours. 01:13:07.010 --> 01:13:13.650 And that is equal to 10 to the 6 megawatt hours. 01:13:13.650 --> 01:13:19.050 And we had, from the previous slide, 10 to the minus 11 01:13:19.050 --> 01:13:20.450 per megawatt hour. 01:13:20.450 --> 01:13:32.600 So the accident rate is 10 to the minus 11 accidents 01:13:32.600 --> 01:13:37.730 per megawatt hour times 10 to the 6 megawatt 01:13:37.730 --> 01:13:42.250 hours per terawatt hour. 01:13:42.250 --> 01:13:44.370 And that gives us-- 01:13:44.370 --> 01:13:45.090 what is it? 01:13:48.830 --> 01:13:51.390 10 to the minus 5? 01:13:51.390 --> 01:13:54.570 10 to the minus 5 accidents per terawatt hour. 01:14:01.950 --> 01:14:13.150 And 10 to the minus 5 accidents times 230,000 deaths per 01:14:13.150 --> 01:14:20.770 accident is equal to 2.3 deaths per terawatt hour. 01:14:24.430 --> 01:14:28.050 So now we can put our number on the chart. 01:14:28.050 --> 01:14:29.510 There it is-- 01:14:29.510 --> 01:14:32.510 2.3 deaths per terawatt hour. 01:14:32.510 --> 01:14:36.110 And if we assume the washing away effect of the cesium, 01:14:36.110 --> 01:14:38.990 it goes down all the way to 1.4. 01:14:38.990 --> 01:14:40.450 So is nuclear safe? 01:14:40.450 --> 01:14:42.070 Yes. 01:14:42.070 --> 01:14:45.190 Nuclear is a pretty safe form of energy. 01:14:45.190 --> 01:14:50.160 It's not as safe as they said it was. 01:14:50.160 --> 01:14:52.000 It's a lot worse than that. 01:14:52.000 --> 01:14:57.520 But it's somewhere between hydro and natural gas. 01:14:57.520 --> 01:14:59.160 It's relatively safe. 01:14:59.160 --> 01:14:59.680 Yeah? 01:14:59.680 --> 01:15:01.100 AUDIENCE: [INAUDIBLE] the other-- 01:15:01.100 --> 01:15:03.800 if they got nuclear wrong, what [INAUDIBLE]? 01:15:03.800 --> 01:15:05.340 R. SCOTT KEMP: Oh, thank you. 01:15:05.340 --> 01:15:05.840 Yeah. 01:15:05.840 --> 01:15:09.640 We really shouldn't trust anything on this chart now. 01:15:09.640 --> 01:15:14.240 So what we're going to do, maybe in the next class, 01:15:14.240 --> 01:15:15.840 is we'll look at other calculations 01:15:15.840 --> 01:15:17.300 for these other factors. 01:15:17.300 --> 01:15:17.800 Yeah. 01:15:20.600 --> 01:15:22.920 And why don't I stop there. 01:15:22.920 --> 01:15:24.470 Yeah.